Abstract: In this study, a comprehensive early warning model combining dynamic meteorological data and static geological conditions was proposed to meet the early warning needs of water damage disasters caused by extreme rainfall in the rainy season in mountainous areas in Zhejiang Province. Based on the slope unit divided by the comprehensive method, the model constructs a slope flood and geohazard meteorological early warning model driven by the SHALSTAB model by superimposing the slope potential degree and real-time rainfall monitoring data, and realizes the disaster probability prediction based on the coupling of "dynamic meteorology-static potential". Based on the model, an early warning system for water damage on the slope of a long-distance natural gas pipeline was developed, and an empirical study was carried out in a mountainous section of Zhejiang Province: by inverting the rainfall, potential distribution and disaster probability spatial characteristics of historical disaster points, the consistency between the prediction results of the model and the actual disaster data was verified to be 57%, and it was confirmed that the early warning system has a certain prediction ability and can be applied to the pipeline operation and management of enterprises. In addition, the system can provide refined management support from surface to point for daily pipeline inspection, which helps to optimize the allocation of inspection resources.