Aims Our objective was to simulate and predict the impact of climate change on potential distribution in the relic and endangered Amygdalus mongolica, hence providing scientific basis for understanding the evolution and protection of this species. Methods Maximum entropy (MAXENT) model was employed to simulate, forecast, compare, analyze, and reveal the changes in the distribution range and spatial pattern in the relic and endangered A. mongolica at the Last Glacial Maximum (based on Community Climate System Model and the Model for Interdisciplinary Research on Climate), and under the historical (1961-1990) and future climate conditions (2020, 2050 and 2080, all based on the Intergovernmental Panel on Climate Change (IPCC) and Special Report on Emissions Scenarios A2A). The accuracy evaluation of repeat models of A. mongolica, and the average probability of occurrence and standard deviation based on repeat models were analyzed with the spatial analysis methods in the ArcGIS10.0. Important findings The potential distribution of A. mongolica under the historical climate conditions centered in Omnogovi and Dornogovi of Mongolia, Bayannur City, Alxa Zuoqi, Ordos City, and western Xilin Gol Meng of Nei Mongol, the central and eastern regions of Hexi Corridor, the northern Ningxia and Shaanxi, and part of the northern Heibei. Furthermore, the distribution of A. mongolica at the Last Glacial Maximum based on Community Climate System Model climate scenario experienced the widely southward shift and range retraction. Last but not the least, under the future A2A climate scenario of IPCC, the potential distribution of A. mongolica would be significantly increased by 2020, and then decreased by 2050, with a slightly increasing trend until 2080. The distribution patterns of A. mongolica showed a large spread and shift to eastern Hebei and the eastern Nei Mongol of China, and to the eastern, northern, and western Mongolia.