The Zoige (Ruoergai) Wetland, China's largest high-altitude peatland region, is crucial for maintaining the water resource balance of the Yellow River. However, the lack of long-term, high-resolution, and densely sampled wetland datasets has impeded a thorough understanding of the wetland's landscape evolution. To elucidate the spatio-temporal evolution trends and driving factors of the Zoige wetland and enhance its protection and management, we concentrated on the marshes and water bodies within the Zoige Wetland, utilizing the Google Earth Engine cloud platform to process Landsat-5/7/8 remote sensing satellite data. Employing random forest classification and sample transfer techniques, we created a wetland spatial distribution dataset covering 1986 to 2024, with a spatial resolution of 30 m and a temporal resolution of one year, achieving an overall classification accuracy exceeding 95.49%. The findings show that by 2024, the total wetland area of marshes and water bodies had diminished by 77.15 km² from 1986, indicating an overall degradation rate of 2.5%. Marshes followed a "growth-decline-growth" trajectory, with key transition points identified in 1996 and 2013. Water bodies exhibited greater variability, following a "decline-growth" pattern, with their area in 2024 being approximately 1.6 times that of 1986. Marsh degradation was closely linked to rising temperatures during the growing season, while increases in precipitation and runoff facilitated the expansion of water bodies. Recent wetland restoration efforts have played a significant role in reversing these degradation trends. However, the connectivity of water bodies and the aggregation of marshes have shown a declining trend. Consequently, future wetland conservation and restoration initiatives should prioritize enhancing the connectivity and aggregation of wetlands. This study provides critical insights and reference data for the protection, restoration, and sustainable management of the Zoige Wetland.