Tongren City of Qinghai Province was selected as the study area, with Sentinel-1A satellite ascending and descending SAR data as the main information source, and SBAS InSAR technology was used to retrieve and extract regional surface deformation information. In combination with the background data of topography, geological structure, meteorology and hydrology, vegetation coverage, historical disasters, etc. in the area, GF-2 optical image screening and confirmation are used to realize the early identification of landslide hazards in the area, and a total of 19 landslide hazards are identified, Among them, 8 are known landslide hazard points and 11 are newly identified hidden danger points. The deformation inversion and hidden danger identification results are analyzed for the accuracy of deformation rate and the difference analysis combined with geometric distortion, the accuracy and effectiveness of the deformation inversion results are evaluated, and the comprehensive and accuracy of the hidden danger identification results are proved. Three typical landslide hazards are selected to analyze the characteristics from the aspects of temporal and spatial distribution characteristics of deformation, optical image characteristics, field deformation characteristics, etc., and comprehensively grasp the hazard preparation conditions and deformation trend of landslide hazards in the area.