This study aims to monitor the ice lakes in the Selin Co Basin using a Composite Index (CI) derived from remote sensing. Ice lakes, important water bodies in plateau regions, are significantly affected by climate change, and their area variations are crucial indicators of environmental and ecological changes, especially under the context of global warming. Traditional remote sensing indices, such as NDWI (Normalized Difference Water Index) and NDSI (Normalized Difference Snow Index), have certain limitations in accurately distinguishing between water and ice surfaces. To address this, a ratio-based composite remote sensing index (CI) is constructed by calculating the ratio of NDWI to NDSI, which enhances the differentiation between water and ice. First, pixels with NDWI and NDSI values less than zero are excluded to ensure the accuracy of the valid areas. Then, a Partial Least Squares (PLS) regression method is applied to determine the classification threshold for CI, which is set to 0.37. CI values between 0 and 0.37 indicate water, while values above 0.37 represent ice. The results demonstrate that the CI effectively improves the accuracy of water-ice classification, providing a more reliable method for ice lake monitoring. Finally, by utilizing Sentinel-2 data from the winter of 2024, this study achieves dynamic monitoring of the freezing conditions of ice lakes in Selin Co, offering valuable data for future climate change research, ecological restoration, and water resource management.