This paper focuses on utilizing Sentinel 2 MSI datasets to generate satellite-derived bathymetry (SDB) maps at a resolution of 10 m for two temporally varying datasets of the study region of Ameland Inlet, located in The Netherlands, by using support vector regression (SVR) technique. The relative performance of Landsat 8 OLI (30 m) datasets with SVR technique is also assessed to demonstrate the complementary nature of these freely available medium-resolution imageries. Further, the root mean square error and mean absolute error between the retrieved and measured bathymetries are estimated and reported to evaluate the capability of SVR in estimating depths. It is evident that the SDBs thus generated using this machine learning approach provide dependable estimations of depths that can further be utilized for various coastal engineering studies.