Abstract: Carrying out researches on the prediction of the temporal and spatial distribution of glacial lakes is of great significance to preventing disasters caused by glacial lake outbursts.Based on the results of GEE supervised classification, this paper analyzed the spatiotemporal variation characteristics of Kunlun Mountain glacial lakes in 2000～2020,explored the applicability of three land use/cover prediction models: multi-criteria CA-Markov, FLUS(Future Land Use Simulation),PLUS(Patch-generating Land Use Simulation Model).Then we obtained the most suitable model for Kunlun Mountain glacial lake prediction, and predicted the future glacial lake change trend in the study area.The results showed that:(1) In 2020,the number and area of glacial lakes in the Kunlun Mountains increased by 39.25% and 81.35% respectively compared with that in 2000,and the glacial lakes distributed more in west and less in east.(2) Among the three prediction models, the multi-criteria CA-Markov Model performed the best in accuracy evaluation, with a Kappa coefficient of 0.86.(3) The area and number of glacial lakes in 2025 will increase by about 6.54% and 9.73% respectively compared with that in 2020.The area and number of glacial lakes in 2030 will increase by about 14.59% and 16.55% respectively compared with that in 2020,and the overall numbers of Kunlun Mountain glacial lakes in the future will still be in a augment trend, however the differences between the east and west areas will decrease.