Sweden has the most islands of any country, with 267,570 counted in its national island register, of which only about 1,000 are inhabited. Finland ranks second at roughly 178,000, and Norway third at about 240,000 depending on the counting method. The ranking hinges entirely on how an island is defined, because no global authority sets a minimum size.
How an island gets counted
The mechanism behind these numbers is a national land registry, not a satellite census. Sweden's mapping authority, Lantmäteriet, maintains a register of every landform that qualifies as an island (ö) under Swedish law. Each polygon of land surrounded by water is recorded with coordinates, area, and a unique identifier, so the count is a byproduct of property mapping rather than a deliberate attempt to win a global contest.
The definition threshold is where rankings break apart. Sweden counts islands down to roughly 25 square meters (about 270 square feet), and even smaller rocks in some datasets. Norway's register includes skerries and outcrops at roughly 10 square meters in some provinces. Finland's count varies by source between about 80,000 and 188,000 islands depending on the minimum area used and whether lake islands are included.
A country with a single cutoff of one square kilometer would produce a short list dominated by large states. A country counting every emergent rock produces a list dominated by archipelagic nations. Neither list is wrong; they answer different questions. The professional caveat is that cross-country comparisons are only valid when the same threshold is applied, and almost no global dataset does that.
Plain-language version: the winner depends on how small a rock still earns the name island.
Quantified context: the numbers behind the podium
Sweden's 267,570 figure comes from Lantmäteriet's register and is the most-cited national total in the world. The number reflects a coastline that includes thousands of kilometers of brackish archipelago in the Baltic Sea, plus tens of thousands of lake islands in the interior. Sweden has about 96,000 lakes larger than one hectare, and many contain islands of their own.
Finland's official line is 188,000 lakes and about 178,000 islands, though the Finnish Environment Institute has published island counts as low as roughly 80,000 under stricter area rules. The discrepancy of more than 100,000 islands in a single country is a warning label on the entire comparison.
Norway holds roughly 240,000 islands and skerries by its national mapping, and some datasets place it above Sweden. Canada counts about 52,000 islands officially, though its Arctic archipelago is immense and its smaller features are poorly surveyed. Indonesia has over 17,000 named islands, with a 2022 government verification placing the official count near 17,000 after years of registry work.
The Philippines reports about 7,641 islands, revised upward from 7,107 in 2017 after better mapping of the Visayas and Mindanao regions. Japan's official count rose to 14,125 in 2023 when the Geospatial Information Authority recounted with a one-hundredth-square-kilometer threshold, which is more than double the previous figure of 6,852. That single redefinition added over 7,000 islands to Japan's tally in one year.
Comparison table: country island counts
| Country | Reported islands | Counting threshold | Inhabited islands |
|---|---|---|---|
| Sweden | 267,570 | About 25 m² | About 1,000 |
| Norway | About 240,000 | About 10 m² (varies) | Several thousand |
| Finland | 178,000 to 188,000 | Varies by register | About 500 |
| Canada | About 52,000 | Varies by province | A few hundred |
| Indonesia | About 17,000 | Named and verified | About 6,000 |
| Philippines | 7,641 | Emergent land above high tide | About 2,000 |
| Japan | 14,125 | 100 m² or larger | About 400 |
Notice that the ranking is not monotonic with coastline length. Canada has the longest coastline in the world at roughly 202,000 kilometers yet trails Sweden by more than 200,000 islands, because its register counts fewer small features. This is the counterintuitive detail: coastline size does not predict island count when registers disagree on what counts.
Edge cases: lakes, tides, and moving land
Lake islands count in Sweden and Finland, which inflates their totals enormously. Sweden's Lake Vänern alone contains thousands of islands. A strict marine-only tally would drop Sweden's lead substantially, though it would probably still rank near the top.
Tidal flats create legal ambiguity. In the Philippines and Indonesia, land that emerges only at low tide may or may not count as an island depending on the register. The 2017 Philippine recount added hundreds of features by clarifying this boundary, which shows how a regulatory decision, not a geological event, can change a national statistic.
Post-glacial rebound moves the land itself. The Baltic region rises about 4 to 8 millimeters per year as the crust rebounds from the last ice age. Over decades, shallow straits close and one island becomes two, or two islands merge into a peninsula. Sweden's island count is therefore not fixed; it drifts with geology and with the mapping office's update cycle.
River islands are another edge case. Bangladesh's braided rivers shift channels constantly, creating and erasing islands (chars) within a single generation. No stable national total is possible there. Any list that treats island counts as permanent facts is quietly wrong.
Misconceptions worth correcting
The most common error is assuming the leaderboard is settled. Wikipedia's list of countries by number of islands has been edited and re-sourced repeatedly, and different versions have placed Sweden, Finland, Norway, and Indonesia at the top. The instability comes from the sources, not from the geography.
A second misconception is that more islands means more territorial sea. Under the United Nations Convention on the Law of the Sea, an island generates a territorial sea of up to 12 nautical miles and an exclusive economic zone of up to 200 nautical miles, but a rock that cannot sustain human habitation or economic life gets no EEZ. Counting rocks therefore does not expand sovereignty. This legal distinction, established in Article 121, is why some states care intensely about which features qualify as islands and others do not.
A third error is treating the counts as comparable. Japan's 2023 jump from 6,852 to 14,125 was a measurement change, not a discovery of new land. Any article that presents these figures side by side without thresholds is misleading.
Practical actions for using these numbers
State the definition before the count. Give the minimum area and whether lake islands are included whenever you cite a national total. Cite the national mapping authority, not aggregator sites. Lantmäteriet, the Geospatial Information Authority of Japan, and Indonesia's Geospatial Information Agency publish the primary registers. Treat rankings as provisional. Registers update on multi-year cycles, so a 2020 figure may be obsolete. Check the legal status separately. An island count is a mapping fact; territorial rights under UNCLOS are a legal question with different criteria. Compare only like with like. If one country uses 25 square meters and another uses 100, do not place them in the same table without a footnote.
For journalists and researchers, the safest sentence is: "Sweden reports the highest count under its national register, which includes features as small as roughly 25 square meters." That sentence survives scrutiny because it names the source and the threshold.
Timeline table: how the counts changed
| Year | Event | Effect on counts |
|---|---|---|
| 1950s-1970s | National mapping agencies digitize coastlines | First machine-countable registers |
| 1990s | Sweden's property register becomes comprehensive | 267,570 figure stabilizes |
| 2017 | Philippines recounts with clarified rules | 7,107 rises to 7,641 |
| 2022 | Indonesia verifies island registry | Official count settles near 17,000 |
| 2023 | Japan recounts at 100 m² threshold | 6,852 rises to 14,125 |
| Ongoing | Baltic post-glacial rebound | Slow, continuous change in Nordic counts |
The pattern is clear: most large jumps come from methodology, not from nature. The 2023 Japanese recount is the cleanest example, adding more than 7,000 islands without a single new island forming.
Limitations and what the numbers cannot tell you
No global body maintains an island census, so every cross-country comparison is an assembly of national registers with different rules. The United Nations Statistics Division tracks land area and coastline but not island counts. The CIA World Factbook lists island counts for some countries but not systematically. This means the ranking is a best-effort synthesis, not an official statistic.
The counts also exclude what many people mean by island. Seasonal sandbars, mangrove clusters, and artificial islands are handled inconsistently. Dubai's Palm Jumeirah and other constructed islands sit outside most national registers entirely.
Finally, island counts say nothing about habitability, biodiversity, or economic value. Sweden's 267,570 islands include roughly 1,000 inhabited ones, a ratio of about 267 to 1. The number is a mapping artifact as much as a geographical fact, and it should be read as one.
If you need a single defensible answer, use Sweden with the threshold attached. If you need a comparison across countries, state the definition for each entry and accept that the leaderboard is provisional until a global standard exists.