Under the rapidly warming climate in the Arctic and high mountain areas, permafrost is thawing, leading to various hazards at a global scale. One common permafrost hazard termed retrogressive thaw slump (RTS) occurs extensively in ice-rich permafrost areas. Understanding the spatial and temporal distributive features of RTSs in a changing climate is crucial to assessing the damage to infrastructure and decision-making. To this end, we used a machine learning-based model to investigate the environmental factors that could lead to RTS occurrence and create a susceptibility map for RTS along the Qinghai-Tibet Engineering Corridor (QTEC) at a local scale. The results indicate that extreme summer climate events (e.g., maximum air temperature and rainfall) contributes the most to the RTS occurrence over the flat areas with fine-grained soils. The model predicts that 13% (ca. 22,948 km2) of the QTEC falls into high to very high susceptibility categories under the current climate over the permafrost areas with mean annual ground temperature at 10 m depth ranging from −3 to −1 °C. This study provides insights into the impacts of permafrost thaw on the stability of landscape, carbon stock, and infrastructure, and the results are of value for engineering planning and maintenance.