Accurate monitoring is a critical step in evaluating the conservation and management needs of endangered species. We evaluated a low cost, effective survey method for monitoring West Indian manatees (Trichechus manatus manatus) in Belize, Central America. The objectives for this paper are (1) to evaluate a count-based population index derived from a boat-based survey method, (2) to examine trends in manatee abundance in the Drowned Cayes area, (3) to conduct a power analysis to explore our ability to detect a trend and the ramifications of survey structure on trend detection. We used a generalized linear model to evaluate the impact of environmental conditions on sighting probability and to determine whether the number of manatees observed per 20-min scan changed from 2001 to 2007. We used simulations to determine statistical power – the ability to detect potential declines of 10%, 25% or 50% over 15 years and for various sampling regimes. The number of manatees sighted per scan was not affected by sighting conditions. There was no change in the mean number of manatees sighted per scan from 2001 to 2007. Our ability to detect a trend ranged from 9% to 100% depending on the level of decline, scan duration, number of points surveyed and number of surveys. This survey protocol is a practical and repeatable way to examine population trends of sirenians in similar habitats around the world.