Species distribution models have become widespread in benthic ecology; however, they have mostly been applied to marine and lentic systems. In this study, we applied a maximum entropy model by using remote sensing-derived environmental variables to predict the distribution of 4 major benthic communities dominated by Tubificidae, Naididae, Echinogammarus ischnus, andDreissena spp, respectively, in the lower Niagara River, NY, USA. The model showed very good accuracy for benthic communities with a narrow distribution range (Tubificidae and Naididae) indicated by the area under the curve test values of 0.906 and 0.987, respectively. In contrast, the model showed poor performance for E. ischnus and Dreissena indicated by the low area under the curve values of 0.615 and 0.618, respectively. Both communities are known to cope with a wide variety of environmental factors and habitats, making their accurate predictions difficult using presence-only data. Our results can further be used to locate important feeding grounds for higher trophic levels, to assess the potential spread of exotic species, and to identify areas for restoration.