This paper deals with a new method for remote sensing data recovery, which is based on the Digital Elevation Model (DEM). In this technique, ground topography acts as the key factor, and random effects of satellite remote sensing are ignored. First, a DEM is created according to the digital contour lines from the basic geographic data. Then, through combining the ground real data with DEM and corrected remote sensing data, the slope & aspect analysis of pixel in remote sensing data is performed, thus building up a model which shows the relationship between the DEM and the remote sensing data so as to characterize the action mechanism of the topography on the remote sensing data by mathematical statistics. Finally, the remote sensing information recovery is conducted for each pixel in remote sensing data. With this method, an excellent information recovery result has been achieved. The technique can also eliminate or reduce the influence of topography on remote sensing data. Therefore, it enhances the practical application of remote sensing techniques under the complex topographic conditions of mountain areas.