We used aerial survey and corresponding digital land cover data to develop species-habitat models to describe breeding-ground distributions and landscape-level habitat associations of greater white-fronted geese (Anser albifrons), Canada (Branta canadensis) and cackling geese (B. hutchinsii), tundra swans (Cygnus columbianus), king eiders (Somateria spectabilis), and long-tailed ducks (Clangula hyemalis). We then used habitat associations in the Queen Maud Gulf Migratory Bird Sanctuary and the Rasmussen Lowlands, Nunavut, Canada, in models to predict distributions of focal species in each study area. We used the receiver operating characteristic (ROC) method and the area-under-the-curve (AUC) metric to evaluate predictive accuracy (hereafter, quality) of models. In the Queen Maud Gulf, AUC values suggested reasonable model discrimination for white-fronted geese, Canada geese, and tundra swans (i.e., AUC > 0.7). Quality of species-habitat models for king eiders and long-tailed ducks was less than other species considered, but these models still predicted encounters and non-encounters significantly better than the null model. For all species, quality of species-habitat models was lesser for the Rasmussen Lowlands than for the Queen Maud Gulf, although discrimination ability for Rasmussen Lowland distributions remained significantly better than corresponding null models for geese and swans, but not for seaducks. Our research suggested that species' distributions modeled with landscape-level habitat data is a tractable method to 1) identify habitat associations, 2) determine key habitats and regions, and 3) predict probable summer distributions of some waterfowl species over relatively large areas of the arctic from satellite imagery. © 2013 The Wildlife Society.