Habitat selection is a critical aspect of a species' ecology, requiring complex decision‐making that is both hierarchical and scale‐dependent, since factors that influence selection may be nested or unequal across scales. Elk (Cervus canadensis) ranged widely across diverse ecoregions in North America prior to European settlement and subsequent eastern extirpation. Most habitat selection studies have occurred within their contemporary western range, even after eastern reintroductions began. As habitat selection can vary by geographic location, available cover, season, and diel period, it is important to understand how a non‐migratory, reintroduced population in northern Wisconsin, USA, is limited by the lack of variation in topography, elevation, and vegetation. We tested scale‐dependent habitat selection on 79 adult elk from 2017 to 2020 using resource selection functions across temporal (i.e., seasonal) and spatial scales (i.e., landscape and home range). We found that selection varied both spatially and temporally, and elk selected areas with the greatest potential to influence fitness at larger scales, meaning elk selected areas closer to escape cover and further from “risky” features (e.g., annual wolf territory centers, county roads, and highways). We found stronger avoidance of annual wolf territory centers during spring, suggesting elk were selecting safer habitats during calving season. Elk selected habitats with less canopy cover across both spatial scales and all seasons, suggesting that elk selected areas with better access to forage as early seral stage stands have greater forage biomass than closed‐canopy forests and direct solar radiation to provide warmth in the cooler seasons. This study provides insight into the complexity of hierarchical decision‐making, such as how risky habitat features and land cover type influence habitat selection differently across seasons and spatial scales, influencing the decision‐making of elk. Scale‐dependent behavior is crucial to understand within specific geographic regions, as these decisions scale up to influence population dynamics.