Abstract: Malus baccata (L.) Borkh. is an important germplasm resource,with high ornamental, economic, and ecological value. In this study, we used the ENMeval packet to test the complexity and performance of the MaxEnt model under different regulation multipliers and feature combinations and chose a fine-tuned setting with the lowest complexity. Based on 602 modern distribution records and eight environmental variables, we simulated the potential distribution of M. baccata during the Last Glacial Maximum, mid-Holocene, present, and 2070s (RCP 8.5). Percentage contribution, permutation importance, and Jackknife tests were used to evaluate the importance of environmental variables. Results showed that when the RM value was 2 and FC was LQHPT, the complexity and degree of overfitting of the model were relatively low. The prediction results of the MaxEnt model were excellent, as indicated by the AUC value of 0.9272 ±0.0019. Under current conditions, the highly suitable region for M. baccata covered Mt. Taihang, Mt. Guancen, and Mt. Lüliang of Shanxi Province, northeast Liaoning and Jilin, Shaanxi, Gansu, and Ningxia junction, northern Hebei, and central and southern Shandong. During the Last Glacial Maximum, the highly suitable distribution region in the north disappeared and moved southward. In the mid-Holocene, the potential distribution regions were similar to those of the present-day, though somewhat contracted to high altitude areas. By the 2070s, the distribution in China will sharply contract to high altitude areas. Both temperature and precipitation factors restricted the potential geographical distribution of M. baccata, with the latter being more important.Predicting how climate change will affect geographic range can provide a scientific basis for the protection and management of the germplasm resources in M. baccata.