Permafrost is a soil medium which is extremely sensitive to climate change, so climate change has an important impact on permafrost. In this study, based on altitude-response model, the elevation data (DEM), longitude data, latitude data, mean annual air temperature data (MAAT) and the vertical lapse rate of air temperature (VLRT) are used to derive permafrost distribution in each decade on the Qilian Mountains over the past 40 years. For accuracy of simulation result, all these data except VLRT were transferred into a raster format and stored by WGS-1984 geographic projection and the same spatial resolution (0.001°, about 100 m). VLRT data can be obtained mainly based on the results of previous studies, which shows that the temperature decreased 5.5° by an increase of each 1000 m in elevation in the Qilian Mountains. Through analyzing, MAAT has a close relationship with altitude, longitude and latitude respectively. As for the MAAT data, linear multiple regression analysis between MAAT, DEM, latitude and longitude data was carried out for each decade. The simulation results are compared based on the Map of Snow, Ice and Frozen Ground in China, the interpretation results including 93 key permafrost point landforms can be used to test the accuracy of the simulation results, which were mainly based on field survey data, remote sensing image and DEM data for manual compilation of the Geomorphologic Atlas of People's Republic of China (1:1,000,000). The overall simulation results have a high accuracy, which can reach more than 80%. Using the altitude- response model, the areas of simulated permafrost distribution on the Qilian Mountains in the 1970s, 1980s, 1990s, 2000s, are 9.75*10~4 km~2, 9.35*10~4 km~2, 8.85*10~4 km~2, 7.66*10~4 km~2, respectively. The permafrost distribution area shows a decreasing trend over the past 40 years. Through the further analysis of the simulation results, from the 1970s to the 1980s, 1980s to 1990s, 1990s to 2000s, the rate of the permafrost loss was 4.1%, 5.3% and 13.4%, respectively, which shows an increasing trend and there exists a leap growth from the 1990s to 2000s. This research can provide an effective method for a better understanding of the dynamic changes of alpine permafrost distribution on the Qilian Mountains for a long series period.