Abstract: As the "Water Tower of Asia", the Qinghai-Tibet Plateau is the birthplace of more than ten major rivers in Asia, and the changes of its water resource environment have a great impact on the downstream ecological environment. The use of remote sensing image classification technology can realize the rapid extraction and monitoring of plateau lake water bodies. According to the characteristics of large area of permanent and semi-permanent snow on the Qinghai-Tibet Plateau, the traditional methods of water body extraction using normalized water body index (NDWI) and supervised classification are usually difficult to distinguish lake water bodies from snow. To address this problem, a decision tree classification model for lake water extraction on the Tibetan plateau is constructed by analyzing the geomorphological environment of the plateau snowpack and lake distribution area using NDWI method and slope elements, and the extraction accuracy of the constructed decision tree is verified by setting up a comparison experiment. The experimental results show that the classification accuracy of the decision tree can reach 94.5%, which is better than 82.5% of the extraction accuracy of NDWI method and 88.5% of the extraction accuracy of supervised classification method.