This study combines ICESat-2 and multispectral imagery data, using satellite-based bathymetry (SDB) methods to estimate the water depth of supraglacial lakes on the Greenland ice sheet. Considering reflectance differences at varying depths, we performed spectral stratification based on these variations to derive more accurate SDB inversion results. This study, with its low cost, spatiotemporal advantages, and higher accuracy, can provide insights into terrain and climate change in the Arctic.