The main goal of the present study is to use satellite data to extract bathymetry maps of coastlines and especially the shores of the Caspian Sea. For this purpose, the area between the Neka power plant and Amirabad port in Mazandaran province was selected as a pilot. Landsat-OLI satellite image was used to extract the bathymetry map of the study area. Simultaneously with the path of the satellite, about 2700 points from the depths of 2 to 11 meters of the Caspian Sea, was measurement, of which 500 points were used as control points and the rest as training samples. The polynomial linear regression method was used to extract the bathymetry map. Also, a stepwise regression method was used to identify the best regression model and select the best independent variables to estimate the depth. Comparison between the water depth map extracted from the Landsat-OLI satellite image with the control points showed that the RMSE value of this sensor in estimating the coastal water depth was about 0.4 m with an average standard error of 7.6%. By considering the turbidity and roughness of the seawater of Caspian Sea, the obtained result is an acceptable accuracy.