Identification of landslide should first understand what it is. Broad sense of landslide means all the slope gravitational erosion phenomena including rock fall, landslide, clastic flow and debris flow and so on; narrow sense of landslide means the phenomenon that parts of slope move searing along one or several surfaces internal the slope under the function of the gravitation. Each type of landslide has own special morphological characteristics on the ground and basic geological environmental conditions and trigger factors for growing. According to the special characteristics the landslide can be identified. Digital Landslide Technique suggested based on the research and practice of RS+GIS for landslide at home and abroad identification of the landslide can be divided into 2 steps roughly. First, with RS+GIS technique, ground features at different time were showed on digital images with different solutions. Through registration and combination with geological structure and geographic information, the base for remote sensing interpretation is established. Second, under the guide of theory of landslide, by means of men-PC interactive mode, the remote sensing interpretation and time-space analysis are carried out and the key elements information of landslide for disaster prevention then are obtained. That methods has not been reached automatic identification of landslide, but the process of establishing the base for remote sensing interpretation has been accomplished with several soft program procedure by PC, so it is considered that the first two steps, digitalization and preprocess for landslide pattern recognition have already realized by PC. Now we should explore the methods of landslide recognition and analysis procedure with computer calculation instead of men-PC interactive mode. For the narrow landslide sense, the landslide topography recognition has already been conducted by computer based on DEM. If the relation between landslide scarp, deposit body on the ground and landslide surface, bed under the ground could be found, their spectral property could be determined and the calculating mode could be established, the landslide pattern recognition based on the remote sensing technique then could be realized.