Synthetic Aperture Radar(SAR),as an advanced remote sensing technology,is widely used in environmental monitoring,resource investigation,and disaster assessment,where it has demonstrated remarkable effectiveness. SAR image change detection is of great significance because it enables rapid and comprehensive damage assessment by analyzing multi-temporal intensity images acquired before and after earthquake-induced landslides. However,excessive noise and insufficient exploitation of image information in SAR data can adversely affect the accuracy of change detection. To further improve detection accuracy,this study constructs difference images based on variations in SAR intensity and applies an adaptive Gaussian thresholding algorithm to determine the optimal threshold. A majority voting method is then employed to extract preliminary landslide candidate areas,which are further refined by incorporating slope values derived from a Digital Elevation Model(DEM)as terrain constraints. Experimental results demonstrate the strong potential of this approach for landslide detection. Under complex terrain conditions,the proposed model successfully identified most landslide events when validated against field-based landslide inventory data,achieving an overall accuracy of 90.7% and showing robust performance. This method provides reliable technical support for post-earthquake disaster assessment and emergency response.