Subtropical reservoirs are susceptible to the eutrophication process, largely due to inappropriate land use and occupation in the watersheds, resulting in changes in the composition of the phytoplankton community, according to the sensitivity and environmental tolerances of the different species. Application of the Reynolds functional groups (RFGs) method to the multispecies phytoplankton community is a valuable approach that can be used to answer ecological questions at different spatial and temporal scales. The Itupararanga reservoir, in the upper Sorocaba River basin (São Paulo state, Brazil), is used to supply drinking water to approximately one million people, as well as for agricultural and recreational purposes. This study examines the relationships between environmental variables and the composition of the phytoplankton community, identifying the main functional groups (RFGs) along a spatial gradient in the reservoir (from upstream to downstream), during one year. Five samplings were performed at seven points in the reservoir, in the rainy and dry seasons (from 2016 to 2017). The effects of spatial heterogeneity influenced the formation of different biomass gradients for each functional group. The most abundant RFGs were Sn (warm mixed environments), H1 (eutrophic, both stratified and shallow lakes with low nitrogen content), and Lo (deep and shallow, oligotrophic to eutrophic, medium to large lakes), which corresponded to the spatial heterogeneity of the reservoir. The variables with the greatest influence on the phytoplankton composition in the reservoir were total nitrogen, nitrite, and the N/P ratio, with phosphorus as the limiting nutrient. The presence of Cyanobacteria (codons Sn and H1) was driven by the stability of the water column, due to the long residence time of the water in the reservoir. This research demonstrated the effectiveness of using RFGs as environmental descriptors, providing an essential tool in aquatic ecology studies.