The Miyun Reservoir, located in the Miyun District of Beijing, China, is the largest comprehensive water conservancy project in northern China and an important ecological protection area. The combined effects of many factors produce ecosystem changes in the basin; thus, it is important to analyze the spatial and temporal changes that occur here. Based on multi-source satellite remote sensing data, we analyzed changes in water body area, water level height, and water storage in the Miyun Reservoir from 2013 to 2022 and determined whether these changes were natural or caused by human activity. As traditional water body area extraction methods can misidentify buildings and mountainous areas as water bodies, we fused multiple deep learning models (U-Net and SegNet) using the adboost method, which combined the advantages of the basic models and achieved an overall recognition accuracy of > 90 %. Using annual variations in water storage at the reservoir, we determined that the water body area increased to 157.58 km² between 2013 and 2022, nearly doubling in size, which corresponded to decreases in cultivated land and vegetated areas. Cultivated land is the main land use type affected by water body erosion. The overall water level height exhibited an upward trend (cumulative increase of 14.8 %), eventually reaching 146.11 m. The water storage volume also increased over time, with a cumulative increase of approximately 436 million m³. On this basis, the influences of temperature, precipitation, and human activity on the spatial and temporal variability of the Miyun Reservoir basin were analyzed. The findings have important implications for global change research within and outside the ecosystem.