Semiarid environments are characterized by the uneven spread of rain throughout the year. This leads to the establishment of a biota that can go through long periods without rain. In order to understand the dynamics of rhizosphere microbial communities across these contrasting seasons in Caatinga, we used the Ion Torrent platform to sequence the metagenome of the rhizosphere of a native leguminous plant (Mimosa tenuiflora). The annotation indicated that most abundant groups detected were the Actinobacteria and Proteobacteria, and the dominant functional groups were carbohydrate and protein metabolisms, and that in the wet season, the communities carried carbohydrate and amino acid metabolisms.The major differences observed between seasons were higher abundance of genes related to carbohydrate and amino acid metabolisms in the rainy season, indicating that the populations present might be better adapted to a higher abundance of organic matter. Besides, no clear separation of samples was detected based on their taxonomic composition whereas the functional composition indicates that samples from the rain season are more related. Altogether, our results indicate that there is al arge functional stability in these communities mostly due to the selection of features that aid the biota to endure the dry season and blossom during rain.