This article proposed a high-precision map based traffic flow optimization model for unmanned mining road networks, aiming to optimize the flow of unmanned vehicles inside the mine through accurate map information and advanced algorithms, reduce traffic congestion, and improve transportation efficiency. By constructing high-precision maps and using graph theory and machine learning algorithms, the road network traffic flow of unmanned mining vehicles was optimized. The four experimental conclusions in the experimental stage indicated that the research model could effectively optimize the traffic flow of unmanned mining vehicles, and improve the efficiency and safety of mining operations. The effectiveness and practicality of the model were comprehensively validated through four experiments during the research phase. In the first road network efficiency evaluation experiment, the average speed of unmanned vehicles increased from 20 km/h to 30 km/h, and the average travel time decreased from 30 minutes to 20 minutes. In the traffic congestion experiment, the research model reduced the congestion relief time from 45 minutes to 25 minutes under high flow conditions of 150 vehicles per hour. In the dynamic route adjustment experiment, when encountering unexpected events, the optimized response time was reduced from 45 minutes for Event 1 to 30 minutes, and from 40 minutes to 25 minutes for Event 2. In the final safety performance evaluation experiment, the optimized accident rate decreased from 0.8 to 0.3; the number of violations decreased from 120 to 40; the number of emergency braking events also decreased from 90 to 30. From the data conclusion, it can be seen that the traffic flow optimization model for unmanned mining vehicles based on high-precision maps has shown significant effects in improving traffic efficiency, reducing congestion, enhancing dynamic route adjustment ability, and enhancing safety, fully verifying its application value in the process of mining automation and intelligence.