The effects of establishing a gray wolf (Canis lupus) population in Yellowstone National Park were predicted for three ungulate species—elk (Cervus elaphus), mule deer (Odocoileus hemionus), and moose (Alces alces)—using previously developed POP-II population models. We developed models for 78 and 100 wolves. For each wolf population, we ran scenarios using wolf predation rates of 9, 12, and 15 ungulates/wolf/year. With 78 wolves and the antlerless elk harvest reduced 27%, our modeled elk population estimated were 5-18% smaller than the model estimate without wolves. With 100 wolves and the antlerless elk harvest reduced 27%, our elk population estimated were 11-30% smaller than the population estimates without wolves. Wolf predation effects were greater on the modeled mule deer population than on elk. With 78 wolves and no antlerless deer harvest, we predicted the mule deer population could be 13-44% larger than without wolves. With 100 wolves and no antlerless deer harvest, the mule deer population was 0-36% larger than without wolves. After wolf recovery, our POP-II models suggested moose harvests would have to be reduced at least 50% to maintain moose numbers at the levels predicted when wolves were not present. Mule deer and moose population data are limited, and these wolf predation effects may be overestimated if population sizes or male-female ratios were underestimated in our population models. We recommend additional mule deer and moose population data be obtained.