Although the number of traffic accidents in China has been declining year by year, the proportion of major traffic accidents in mountainous areas has continued to rise, reaching 60% in recent years. This shows that a mountain road safety system has not been effectively established. The majority of the methods used in mountain road safety assessments are statistical analysis methods based on accident causation theory. However, these expert methods greatly depend on massive amounts of traffic environmental data, which are very difficult and time-consuming to obtain. This study had the goal of reducing the environmental monitoring data required for this assessment and increasing analysis efficiency. Based on the research of road traffic accident data and traffic flow characteristics in mountainous areas, an attribute reduction algorithm was developed to weigh the significance of the data and construct a systemic analysis mothed. The proposed risk analysis model, which considers the accidents, roadside environment, and traffic safety facilities, is based on Shaanxi province data. The results from the actual implementation of this model indicate that the approach has good operability and holds great promise for mountain road safety assessment.