China is in the process of industrialization with the development and utilization of material energy as the main feature, and at present, metal open pit mining is still one of the high-risk industries in China. With the increase in the intensity of mineral resources development, mine safety situation is deteriorating, the probability and severity of mine accidents are increasing. Open-pit mining is the same as underground mining is the focus of safety, but also difficult. It is an important issue for mine safety to eliminate the hidden dangers of open pit mining, control mine production accidents, and reduce casualties and property losses. In this paper, analyze the internal and external environment of the mine operation, and classify the subsystem of metal open-pit mine operation site, safety management environment and surrounding natural environment for the requirements of safety and production risk management in metal open-pit mine. The risk assessment index system was established by classifying the risk source identification of metal open pit operation site subsystem, safety management environment subsystem and natural environment factor subsystem. Through the cloud modeling method based on probability theory and fuzzy mathematics, the risk assessment index system of Dabaoshan Metal Open-pit Mine was evaluated in terms of risk likelihood and risk severity, and then the overall risk level of Dabaoshan Metal Open-pit Mine and the risk levels of three subsystems were obtained through the risk matrix. The overall risk level of Dabaoshan Metal Open-pit Mine and the risk levels of three sub-systems were obtained through the risk matrix. Combined with the situation of the open pit of Dabaoshan Metal Open-pit Mine and the current situation of safety management, the graded risk management decision measures will be analyzed. Then, sensitivity analysis will be conducted for the four main evaluation units, namely slope damage, blasting injury, mining and transportation operation injury and discharge site hazard, which are divided into the open-pit metal mine risk assessment subsystem according to the production process. In order to study and analyze the key risk points of this subsystem and the synergy of risks, the subsystem is modeled by Bayesian networks. The key risk paths were obtained from the results of the sensitivity analysis, and targeted risk control measures were proposed based on the risk ranking.