Wetland carbon sink is considered to be one of the most important components of the global carbon cycle. Space-borne remote sensing serves as a vital data source for the classification and carbon sink estimation of wetland. However, inadequate spatial resolution often impedes the accurate classification of different vegetation types, substantially affecting the precision of carbon sink assessments. We propose a framework for the accurate classification and dynamic carbon storage monitoring in co