Resumo. There are different methodologies used to make an inventory of social wasps. In general, these methods are divided into active search and passive collections. Each method has a different performance, depending on the environment in which the collection is being carried out. Thus, the choice for the proper methodology according to the study area will impact the success of sampled species. This study aimed to evaluate the performance of sampling methodologies for social wasps (active search and bait traps) in different phytophysiognomies (Regenerating Cerrado, “Cerradão”, Restored Forest, Semideciduous Forest and Riparian Forest) in the state of São Paulo, Southeastern Brazil. The active search sampled a greater number of species, with the exception of the Regenerating Cerrado. Regarding abundance, the active search was also more efficient, except in “Cerradão”, where there was no difference between the methods, and in Regenerating Cerrado, where the traps sampled a higher number of individuals. The fact that none of the methods collected all the sampled species indicates that it is appropriate to use more than one collection method in order to obtain a sampling closer to the true richness of the studied sites.