Abstract: To reveal the correlation mechanism between heavy rain and road water damage disasters in Hebei Province, this study integrated meteorological and disaster data, and adopted methods such as composite analysis and Random Forest model to explore the impacts of precipitation and terrain on road losses as well as the prediction method. The results show that precipitation and wind fields evolve dynamically during road water damage events, with wind fields dominating water vapor transport and precipitation maintenance; the maximum process precipitation is the core influencing factor (relative importance:44.4%), and precipitation intensity is significantly positively correlated with road losses. The comprehensive index constructed based on Random Forest weights and Pearson correlation coefficients shows a significant correlation between predicted and actual losses (R=0.283, passing the 95% significance test). The eastern foot of the Taihang Mountains is the core disaster zone, while the Zhangjiakou mountainous area has a relatively high loss rate due to fragile terrain; gravel pavements in mountainous areas suffer the most severe losses, asphalt pavements in plains have the optimal damage resistance, and moderate-intensity precipitation is the loss-sensitive interval in coastal areas.This study can provide scientific support for regional highway disaster prevention and mitigation, pavement type selection and risk early warning.