Abstract: The influence weights of the disaster factors were determined to construct a comprehensive hazard index model based on the dominance analysis method with the houses as the disaster-bearing bodies by using the precipitation data of 142 national meteorological stations in Hebei Province and the historical disaster data from 1984 to 2014. Taking the environment vulnerability elements as indicators, Hebei was divided into five regions based on the K-mean clustering analysis method. An exponential function was used to fit the change of house loss and comprehensive index. The thresholds of different grades in each district were then determined and verified by 124 cases from 2015 to 2019. Results showed that the number of housing disasters caused by heavy rain in the north was higher than that in the south, with more than 10 times in the mountainous area. The influence weight of the total rainfall process (68%) was the largest. The weights of the maximum daily rainfall, duration, and maximum hourly rainfall were 22%, 6%, and 4%. The best classification points of simulated disaster were 1180, 335, and 235 damaged houses, and the proportion of the simulated disaster level consistent with the actual disaster level was the highest (67.4%). In the threshold test, the accuracy rate of disaster cases from 2015 to 2019 was 69.8%, and the accuracy rate of the mild disaster level was the highest.