Amazon streams present great biodiversity and offer several ecosystem services, but these systems are threatened by multiple land uses. The changes created by land use are expected to drive the composition of species, ultimately changing the trophic relationships of several biological groups, including zooplankton. We investigated if land use changes the composition of zooplankton functional guilds in Amazon streams and which are the local (physical-chemical) variables driving the zooplankton functional guilds in the land-use gradient. Zooplankton and physical-chemical variables were sampled in 17 water bodies in the municipality of Barcarena, Pará, Brazil in 2018 and 2019, five sampling sites were in the Pará River and 12 in streams. Forest cover (a proxy for land use) was determined through digital image processing and converted in percentage. Zooplankton species were classified into five functional guilds (filter, raptorial, scraper, suctor, and predator feeders). We recorded 98 zooplankton taxa and filters were the most abundant functional guild. The composition of zooplankton functional guilds did not change in the land use gradient. However, the distribution of zooplankton functional guilds in Amazon streams was determined by local environmental variables related to the feeding strategies. Scraper-feeders (cladocerans) were positively related to greater canopy cover, suctor-feeders and predator-feeders (both rotifers) were related to greater total phosphorus, whereas filter-feeders (rotifers, cladocerans, and copepods) and raptorial (copepods) were related to total suspended solids. This study brings new information about zooplankton in Amazon streams that are under-studied. The functional approach clarifies the patterns observed and reflects the trophic relationships in which the zooplankton community is involved in streams under a degree of land use, i.e., scraper-cladocerans can represent more preserved streams under greater canopy cover, whereas the other functional guilds were related to variables that can represent more altered streams.