Coastal bathymetry data are great important environmental resources and disaster management. Today, there are many technologies used to explore the bathymetry, depending on the purpose of surveying. But almost technologies can be time consuming, complicated, and quite expensive. Remote sensing by satellite imagery is one of technology used to estimate the coastal bathymetry in term of accuracy, quality and up to datedness, timely availability and cost effectiveness. Especially the Coastal Blue Band of WorldView-2 for bathymetric measurements will improve both in depth and accuracy. From investigated the relation between surface reflectance and coastal bathymetry of the eight bands of WorldView-2 that band 1,2,3 and 4 are good reflection and when the depth are increase, the digital number (DN) will decrease. The objective of this research is to developing the classification of the coastal bathymetry estimation from satellite imagery. By utilized WorldView-2 satellite images along with regressive function and improving the accuracy result by comparing with in-situ truth depth to assess a coastal bathymetry.