Vector-borne diseases are an increasing issue to public health, endangering billions of people worldwide. Controlling vector mosquitoes is widely accepted as the most effective way to prevent vector-borne disease outbreaks. Mosquito surveillance is critical for the development of control strategies under the integrated vector management framework. We hypothesize that the effectiveness and reliability of using BG-Sentinel traps for the surveillance strongly depend on the bait used to attract mosquitoes. The objective of this study was to compare the effectiveness of BG-Sentinel traps baited with CO2 and BG-Lure. A total of 72 traps were deployed for 48 hours once a week for four weeks. For the initial 24-hour period, the traps were baited with CO2, and then for an additional 24 hours using the BG-Lure. Collected mosquitoes were analyzed using the Generalized Estimating Equation for repeated measures analysis. Biodiversity was assessed by the Shannon and Simpson indices and by individual rarefaction curves and SHE profiles. A total of 5,154 mosquitoes were collected, from which 3,514 by traps baited with CO2 and 1,640 mosquitoes by traps baited with BG-Lure. Aedes aegypti and Culex quinquefasciatus were the most abundant and dominant species. Results from the Generalized Estimating Equation models indicated that more than twice as many mosquitoes were attracted CO2 than to the BG-Lure. The comparison of attractiveness of CO2 and BG-Lure to Ae. aegypti and Cx. quinquefasciatus was non-significant, suggesting that both species were equally attracted by the baits. The individual rarefaction curves for Ae. aegypti and Cx. quinquefasciatus imply that traps baited with BG-Lure underestimated mosquito species richness compared to those baited with CO2. BG-Lure were less effective in attracting mosquitoes with low abundances and failed to collect Cx. coronator and Cx. nigripalpus, which were consistently collected by traps baited with CO2. According to our results, CO2 significantly (P<0.05) attracted more mosquitoes (2.67 adjusted odds ratios) than the BG-Lure when adjusted for time and species, being more effective in assessing the relative abundance of vector mosquitoes and yielding more trustworthy results. Traps baited with CO2 collected not only more specimens, but also more species in a more consistent pattern.