Abstract: Mianluening, located in Qinba mountain area, is one of the areas with the most serious geohazards in Shaanxi Province. This paper adopts 12 factors closely related to the occurrence of geohazards to construct a sample set based on GIS and machine learning technology, and then selects five machine learning algorithms to model the risk of geohazards in Mianluening area. The experiments demonstrated that the Random Forests model is the most suitable one to simulate the occurance of geohazards in the study area. Further division of geohazards risk is then performed by the aforementioned model, which is useful to guide the geohazards investigation and prevention.