：For evaluating the potential of Sentinel imagery for the inversion of Above-Ground Biomass(AGB) of Kunming Dianchi lake wetland, Sentinel SAR and multispectral imagery are used as data sources respectively, and various biomass prediction models are developed through the conventional linear regression and other machine learning algorithms. SAR raw polarisation backscatter data, multispectral bands, vegetation indices, and canopy biophysical variables are extracted. These models have 0.619-0.84 correlation agreement of observed and predicted values, and root mean square error of 40.14-59.7 t/ha. The SAR-based model has the lowest accuracy. Among the Sentinel-2 multispectral bands, the red and red edge bands(band 4,5 and 7), are the best variable set combination for biomass prediction. The model based on the biophysical variable―Leaf Area Index(LAI) derived from Sentinel-2 is more accurate in predicating the overall AGB. The study demonstrates encouraging results in biomass mapping of Dianchi lake wetland by using the freely accessible and relatively high-resolution Sentinel imagery.