BACKGROUND: This study characterized Anopheles mosquito larval habitats in relation to ecological attributes about the habitat and community-level drainage potential, and investigated whether agricultural activities within or around urban households increased the probability of water body occurrence. Malindi, a city on the coast of Kenya, was mapped using global positioning system (GPS) technology, and a geographic information system (GIS) was used to overlay a measured grid, which served as a sampling frame. Grid cells were stratified according to the level of drainage in the area, and 50 cells were randomly selected for the study. Cross-sectional household and entomological surveys were conducted during November and December 2002 within the 50 grid cells. Chi-square analysis was used to test whether water bodies differed fundamentally between well and poorly drained areas, and multi-level logistic regression was used to test whether household-level agricultural activity increased the probability of water body occurrence in the grid cell. RESULTS: Interviews were conducted with one adult in 629 households. A total of 29 water bodies were identified within the sampled areas. This study found that characteristics of water bodies were fundamentally the same in well and poorly drained areas. This study also demonstrated that household-level urban agriculture was not associated with the occurrence of water bodies in the grid cell, after controlling for potential confounders associated with distance to the city center, drainage, access to resources, and population density. CONCLUSIONS: Household-level urban agricultural activity may be less important than the other types of human perturbation in terms of mosquito larval habitat creation. The fact that many larvae were coming from few sites, and few sites in general were found under relatively dry conditions suggests that mosquito habitat reduction is a reasonable and attainable goal in Malindi.