Periodic declines in mule deer (Odocoileus hemionus) populations across much of their range in recent decades, garnered much interest from federal and state wildlife agencies. Effective management of mule deer requires an understanding of how they interact with their environment. Our objectives were to quantify resource use by adult female mule deer during late gestation, provisioning of offspring, and following the loss of offspring. We conducted our study on the Starkey Experimental Forest and Range in northeastern Oregon. We evaluated resource selection by mule deer using a Random Forest machine-learning approach. We assessed temporal variation in resource selection across three time periods: third trimester of pregnancy, 30 days post-parturition, and following the loss of offspring. Mule deer selected for further distances to roads during late stage pregnancy and after the loss of offspring. In contrast, mule deer selected for distances closer to roads during the 30 days post-parturition. We observed higher selection for distances closer to water while rearing young than the other two time periods. Our analysis also incorporated space use by elk (Cervus canadensis). Mule deer showed avoidance for areas with high probability of elk use across all time periods.