Abstract: Taking the collapse area of Chengguan Town in Zhouqu County as an example, through field geological investigation and UAV aerial photography, we established a point cloud model of the slope rock and a three-dimensional terrain model of the slope in the collapse area of Chengguan Town. Based on the point cloud data, we established a structural surface identification and interpretation algorithm to automatically obtain the structural surface yield information and extract geometric information such as structural surface spacing and area. Using the structural surface information and uncertainty methods, we obtained the distribution of volume and shape of the hazardous rock mass. Based on the three-dimensional terrain model, the Unity 3D engine is used to simulate the movement paths, heights, energy, and other elements of different volumes and shapes of falling rocks in the collapse area of Chengguan Town in three-dimensional space. By considering the uncertainty of the parameters affecting the trajectory of the falling rocks, we obtained better simulation results of the three-dimensional movement characteristics of the falling rocks. Finally, the risk assessment of the collapse disaster in Chengguan Town, Zhouqu County, was carried out by combining the simulation results. The results show that most of the avalanche disaster risks in the study area are expressed as low to medium level. Among these, the risks of cubic blocks and slab-like blocks in the C3 area are medium risks; the risks of 5 m³ falling rock blocks in the C3 area are medium risks, and the risks of 17 m³ falling rock blocks in the C2, C3, and C4 areas are medium risks, which pose a more serious threat to the buildings at the foot of the slopes and the people. The results of this study show that it is reasonable to utilize the above method to assess the risk of dangerous rocks on single high steep slopes, which can provide suggestions and scientific references for the prevention and control of related landslide disasters.