Satellite images of shallow water have a potential of enabling high-resolution spatial-temporal monitoring of water depth and bottom feature.However, traditional methods based solely on the spectral information cannot estimate them simultaneously without hyperspectral sensors or additional information on bottom reflectivity.In this paper, the effectiveness of utilizing the textural information (texture features derived from Gray Level Co-occurrence Matrix) as well as spectral information is examined through depth estimation and bottom classification experiments with a Quick Bird image. As a result, textural information is found to be effective in enhancing the accuracy.