<p>Following an earthquake, the landslide resulted in significant social and economic losses. However, as one of the key methods for landslide prediction, landslide susceptibility exhibits a spatial-temporal evolution as the evaluation factors change over time. Therefore, based on terrain factors (elevation, gradient, and aspect) and distance factors (distance from roads, distance from channels, and distance from epicenters), this study selected the spatial-temporal evolution of annual normalized difference vegetation index (NDVI), annual precipitation, and annual average temperature as evaluation factors to provide an accurate and objective landslide susceptibility. In this study, the variation trend models of the Slope model and the Theil-Sen median model were used to evaluate the evolution of selected factors. Based on the principle of ICM-LR and ICM-RF ensemble models, a universal model was constructed by unifying the weight standard of multi-phase evaluation factors, and the spatial-temporal evolution of landslide susceptibility in Shifang City from 2008 to 2022 was analyzed. According to the research results, the area under the curve (AUC) of landslide susceptibility obtained by comparing the four-phase universal model, the Sen model, and the Slope model is greater than 87%. This study not only guides the prevention and control of landslides after earthquakes and local urban planning but also offers a reference value for the study of the temporal and spatial evolution of landslides in earthquake-prone regions worldwide.</p>