Abstract: As an objective system, the evaluation model of weights of evidence can calculate the weight values of different factors, assign the weight process objectively, and distinguish the impact of different factors on the sensitivity of collapse. In order to understand the sensitivity of different factors to collapse, this paper uses the information of the collapse disasters that have already occurred and used the example of the Yan Duhe Town in Badong County, Hubei Province as an example. 106 collapse points are selected, and the weight method of evidence is adopted. The sensitivity analysis of six basic factors is carried out by ArcGIS software. The six factors include slope, slope direction, distance to fault, distance to water system, distance to highway, and formation lithology. 78 randomly selected collapse points are used as a training set for modeling, and the other 28 are used as a test set to test the evaluation result. The analysis results show that the sensitivity of the distance from the collapse point to the highway is the largest. The area with high sensitivity to collapse is linearly distributed near the road and the river. The extremely sensitive area and the high sensitivity area account for 15.27% of the whole research area. There are 22 collapse points in the test set distributed in the area. This is to say that 78.58% of the collapse points are correctly evaluated, which is consistent with the on-site investigation results. The collapse sensitivity map obtained has a guiding role in disaster prevention and reduction work.