Areas identified as winter range are important seasonal habitats for mule deer (Odocoileus hemionus) because they can moderate overwinter mortality by providing thermal cover and forage. Therefore, identifying seasonally important resources is a conservation priority, especially when sensitive areas are proposed for development. We used data collected from global positioning system (GPS) collars fitted on female mule deer (n=19; one location every 3 h) to identify resources important during winter (23 February 2011-***2030 April 2011; 1 November 2011-***2015 January 2012) in a region spanning southern Wyoming and northern Colorado that has been proposed for wind energy development. The study period included portions of two consecutive winters but were pooled for analysis. We used methods to account for GPS biases, fractal analyses to determine perceived spatial scale, and discrete choice models and conditional logistic regression to assess resource selection prior to development (i.e., baseline data). Resource selection by female mule deer revealed similar patterns between active (0600-1800 hours) and nonactive (2100-0300 hours) periods. Deer selected most strongly for proximity to rock outcrops and shrubland and average values of slope. Deer tended to avoid roads and grasslands; all other landscape features had minimal influence on resource selection (hazard ratios near, or overlapping, 1). Using the fixed-effects coefficient estimates, we developed two spatially explicit maps that depicted probability of mule deer occurrence across the landscape. Based on an independent validation sample, each map (active and nonactive) validated well with a greater percentage of locations occurring in the two highest probability of use bins. These maps offer guidance to managing mule deer populations, conserving important seasonal habitats, and mitigating development (e.g., wind energy) in areas identified as important to mule deer.