Abstract: Wetlands are transitional ecosystems formed through interactions between terrestrial and aquatic environments. They are among the most important ecosystems on Earth and provide essential ecological functions, including hydrological regulation, carbon cycling, biodiversity conservation, and water purification. Wuhan, one of the major transportation hubs and economic centers in China, possesses abundant wetland resources. However, rapid urbanization and increasing human activities have exerted considerable pressure on its wetland ecosystems in recent decades. Wetland vulnerability reflects the tendency of wetland ecosystems to undergo degradation or functional loss under external disturbances. Therefore, investigating the dynamic evolution of wetland vulnerability is essential for wetland conservation and regional ecological management. In this study, Wuhan wetlands were selected as the research object. Remote sensing image preprocessing was conducted on the Google Earth Engine (GEE) platform. A total of 43 spatiotemporal variables were constructed from four categories: polarization features, spectral features, texture features, and topographic features. NDVI and NDWI time series were reconstructed using the Harmonic Analysis of Time Series (Hants) algorithm to capture seasonal and phenological characteristics of wetlands. Based on automatically generated sample datasets, the UNet3+ model was employed to extract wetland types in Wuhan from 2016 to 2024. To evaluate urban wetland vulnerability comprehensively, a ‘Pattern-Process-Function-Stress’ (PPFS) framework was established. Principal Component Analysis (PCA) was used to determine indicator weights objectively, and the Jenks Natural Breaks method was applied to classify wetland vulnerability into five levels. The results showed that the areas of lake and river wetlands in Wuhan generally exhibited a recovery trend, and the connectivity of major river systems was significantly improved. However, marsh wetlands in the southwestern region decreased substantially due to agricultural development. These wetlands showed a clear conversion trend toward aquaculture ponds and cultivated land. The Wetland Vulnerability Index (WVI) displayed an overall fluctuating downward trend, indicating a gradual improvement in wetland ecosystem stability. This finding confirms the positive effects of watershed ecological restoration and environmental governance. Nevertheless, the WVI of several important wetlands increased during the study period, suggesting that these areas should remain priorities for future conservation and restoration. The Ecological Pattern Index (EPI) revealed relatively high landscape fragmentation in the central urban area and the Sheshui River Basin. The Ecological Process Index (EPOI) showed a typical spatial pattern characterized by ‘high values in central areas and low values in peripheral areas’. The Ecological Function Index (EFI) gradually became more stable over time. Meanwhile, the Ecological Stress Index (ESI) shifted from concentration in the urban core to a more dispersed distribution across the city, indicating that urbanization pressure has gradually expanded toward suburban regions. Overall, wetland restoration and ecological management in Wuhan have achieved notable progress. However, future wetland governance should shift from area expansion to quality improvement. Greater attention should be given to precise monitoring and ecological restoration in core wetland regions and highly vulnerable areas. These efforts are necessary for building a more stable and resilient urban wetland ecological security pattern. Future research should further optimize sample spatial distribution and improve classification accuracy. In addition, the wetland vulnerability assessment framework should be refined according to the ecological characteristics of Wuhan. This study reveals the spatiotemporal heterogeneity of wetland vulnerability in Wuhan and provides scientific support for wetland restoration and sustainable urban ecological management.