Abstract: Under the " Power Transmission from Xinjiang to Other Regions" policy,the accelerated development of coal resources in Xinjiang has led to ecological degradation. In order to better implement national environmental protection measures,this study employs the Remote Sensing Ecological Index (RSEI) to systematically analyze the ecological and environmental changes from 2003 to 2023 in two representative coal mines in Xinjiang:the Yibei Coal Mine and the Xiheishan Coal Mine. The results reveal the following findings:① the ecological environment of the Yibei Coal Mine exhibited a pattern of " initial degradation followed by improvement," with a significant degree of ecological restoration. In contrast,the Xiheishan Coal Mine experienced a more complex trajectory characterized by " initial improvement,subsequent degradation,and eventual recovery," although the overall improvement was relatively limited. ② RSEI driving force analysis indicated that ecological changes in both mines were jointly influenced by natural and anthropogenic factors. In the Yibei Coal Mine,natural factors such as temperature and precipitation played critical regulatory roles during both the initial degradation and recovery phases,while land use changes emerged as the dominant anthropogenic driver of ecological degradation. Similarly,in the Xiheishan Coal Mine,precipitation and other natural factors significantly influenced ecological changes,particularly during the early and recovery stages. At the same time,land use remained the primary human-induced cause of ecological deterioration. ③ The findings also suggest that different climatic zones necessitate distinct ecological restoration strategies following mineral resource exploitation. Therefore, the results of this study may serve as a valuable reference for ecological management and restoration efforts in coal mines located in semi-humid and semi-arid regions. 

Key words: Yibei Coal Mine,Xiheishan Coal Mine,RSEI,principal component analysis, entropy weight method, geographic detector