Environmental relationships can differ across the geographic range of species, especially for widespread generalists. Because habitat specialists are more vulnerable to environmental changes, incorrect assumptions about consistent habitat associations could hinder strategic conservation efforts. We used species distribution models (SDMs) to evaluate intraspecific variation in habitat associations for a habitat specialist of conservation concern, the pygmy rabbit (Brachylagus idahoensis), which is endemic to the sagebrush biome of the western United States. Our goal was to model habitat associations for pygmy rabbits across a portion of their range to evaluate regional variation and contrast predictions with results from a model developed at the rangewide extent. We created inductive SDMs using maximum entropy methods within five ecological regions that encompassed about 20% of the species rangewide distribution and spanned diverse environmental gradients. We included a suite of environmental predictor variables representing topography, vegetation, climate, and soil characteristics. Results of the regional models identified substantial variation in habitat associations across the five regions, with each retaining a unique set of environmental predictors. Bioclimatic variables were the most influential environmental parameters in all five regions, but the specific variables differed. The models developed at regional extents predicted smaller areas of habitat (an average of 15% less for suitable habitat and 80% less for primary habitat) than predictions generated from a model developed at the rangewide extent. Because bioclimatic variables were effective in discriminating areas used by pygmy rabbits, they also provided an opportunity to assess potential changes in habitat distribution by incorporating future climate projections. Distributions modeled under two mid‐century emission scenarios projected substantial reductions in suitable habitat for pygmy rabbits across most regions and pronounced variation among regions in the magnitude and direction of the climate effects. Collectively, results of this work underscore the need to incorporate regional variation in habitat associations into planning for current and future conservation and management strategies.