The Hindu Kush Himalaya (HKH) is known as the “Water Tower of Asia”, where water influences changes in the global water cycle and ecosystem. It is very important for this region to efficiently measure the status of water and monitor its changes. We proposed a deep learning network based on the state-of-the-art vision Transformer and used it to extract water surface from Sentinel-2 images of the Hindu Kush Himalaya region. This region covers 594 tiles of Sentinel-2 images. Since July to October is the rainy season in the HKH region, we used the images with less than 20% cloud cover during this period in 2021 to extract water surface.The water surface extraction results with different dates from the same tile were composited using a maximum-area algorithm, and the composited results were merged to form the water surface extraction result of the HKH region. The Water Surface product is shared in the form of shpfile.