Ocelots (Leopardus pardalis) are listed as least concern on the International Union for Conservation of Nature (IUCN) Red list of Threatened Species, yet we lack knowledge on basic demographic parameters across much of the ocelot's geographic range, including population density. We used camera‐trapping methodology and spatially explicit capture‐recapture (SECR) models with sex‐specific detection function parameters to estimate ocelot densities across 7 field sites over 1 to 12 years (from data collected during 2002–2015) in Belize, Central America. Ocelot densities in the broadleaf rainforest sites ranged between 7.2 and 22.7 ocelots/100 km2, whereas density in the pine (Pinus spp.) forest site was 0.9 ocelots/100 km2. Applying an inverse‐variance weighted average over all years for each broadleaf site increased precision and resulted in average density ranging from 8.5 to 13.0 ocelots/100 km2. Males often had larger movement parameter estimates and higher detection probabilities at their activity centers than females. In most years, the sex ratio was not significantly different from 50:50, but the pooled sex ratio estimated using an inverse weighted average over all years indicated a female bias in 1 site, and a male bias in another. We did not detect any population trends as density estimates remained relatively constant over time; however, the power to detect such trends was generally low. Our SECR density estimates were lower but more precise than previous estimates and indicated population stability for ocelots in Belize.