Introduction The mosquito Aedes aegypti, vector of dengue, chikungunya and yellow fever viruses, is an important target of vector control programs in tropical countries. Most mosquito surveillance programs are still based on the traditional household larval surveys, despite the availability of new trapping devices. We report the results of a multicentric entomological survey using four types of traps, besides the larval survey, to compare the entomological indices generated by these different surveillance tools in terms of their sensitivity to detect mosquito density variation. Methods The study was conducted in five mid-sized cities, representing variations of tropical climate regimens. Surveillance schemes using traps for adults (BG-Sentinel, Adultrap and MosquiTRAP) or eggs (ovitraps) were applied monthly to three 1 km2 areas per city. Simultaneously, larval surveys were performed. Trap positivity and density indices in each area were calculated and regressed against meteorological variables to characterize the seasonal pattern of mosquito infestation in all cities, as measured by each of the four traps. Results The House Index was consistently low in most cities, with median always 0. Traps rarely produced null indices, pointing to their greater sensitivity in detecting the presence of Ae. aegypti in comparison to the larval survey. Trap positivity indices tend to plateau at high mosquito densities. Despite this, both indices, positivity and density, agreed on the seasonality of mosquito abundance in all cities. Mosquito seasonality associated preferentially with temperature than with precipitation even in areas where temperature variation is small. Conclusions All investigated traps performed better than the House Index in measuring the seasonal variation in mosquito abundance and should be considered as complements or alternatives to larval surveys. Choice between traps should further consider differences of cost and ease-of-use.