Translocation of bighorn sheep (Ovis canadensis) is time, labor, and cost intensive and, therefore, high levels of success are desirable. We tested a widely used habitat suitability model against translocation success and then modified it to include additional factors which improved its usefulness in predicting appropriate translocation sites. The modified Smith habitat suitability model for bighorn sheep was 64% accurate in predicting success or failure of 32 translocations of bighorn sheep into the Rocky Mountains, Colorado Plateau desert, and prairie-badlands of six states. We had sheep location data for 13 populations, and the modified habitat model predicted the areas used by bighorn sheep with greater than 90% accuracy in eight populations, greater than 55% accuracy in four populations, and less than 55% accuracy in one population. Translocations were more successful when sheep were placed into discrete habitat patches containing a high proportion of lambing period habitat (>10% of suitable habitat, p = 0.05), where animals had a migratory tendency (p = 0.02), no contact with domestic sheep (p = 0.02), or greater distance to domestic sheep (>23 km, p = 0.02). Rate of population growth was best predicted by area of lambing period habitat, potential area of winter range, and distance to domestic sheep. We retested the model using these refined criteria and the refined model then predicted success or failure of these 32 translocated populations with 82% accuracy.