Abstract: The wetlands on the eastern Qinghai-Tibet Plateau (EQXP) are a vital global ecological barrier. However, under the dual pressures of climate change and human activities, these wetland systems face environmental challenges such as retrogressive succession, aridification, and desertification. This study utilized the Google Earth Engine (GEE) remote sensing cloud computing platform, and integrated high-resolution imagery, multi-source geoscience datasets, and field survey samples. By employing object-based image analysis (OBIA), logistic regression, and MaxEnt models, we systematically assessed the spatiotemporal dynamics of alpine wetlands in the EQTP and their key driving factors. The results showed that when classifying alpine wetlands using OBIA, the classification accuracy improved, saturated, and fragmented as the image resolution increased from 30 m to 5 m. At a resolution of 10 m (Scale=26), the structure and spatial distribution of marsh wetlands could be accurately identified, with a total wetland area of 174545.6 hm². Wetland distribution was driven by a combination of climatic factors (temperature, precipitation), topographic factors (elevation, slope), and human activities (road density, settlement distribution). The optimal spatial scale for analyzing these driving forces was 500 m (AUC=0.81), confirming that climatic factors dominate long-term changes, while human activities significantly influenced ecological patterns. From 2021 to 2040, the area of highly suitable wetland zones under low-emission scenarios was larger than under high-emission scenarios, with warming causing the contraction of highly suitable zones (very high) to higher altitudes. From 2041 to 2060, the area of excellently suitable wetland zones decreased as regional temperatures rise. From 2081 to 2100, high carbon emission scenarios led to warming in the high-altitude regions of the study area, increasing the suitability for wetland distribution. This study proposes a GEE-based OBIA method for estimating alpine wetland resources, integrating logistic regression and MaxEnt to reveal the spatiotemporal dynamics of alpine wetlands, providing an effective technical framework for wetland research on the Qinghai-Tibet Plateau.