Faced with the huge pressure of new self-driving vehicle technology and relative novel companies, traditional car manufactures are having to prepare for the challenge of the century. In general, the artificial intelligence equipment is high-cost. This research breakthrough uses lower-cost hardware but improves the accuracy of controlling a mobile robot with a high-precision operation model. Therefore, it can help to rapid development for self-driving industry regarding cost-effectiveness. In addition, this technology can be applied to the self-driving car in industry, it can also greatly improve the feasibility of outdoor automatic logistics. Especially it can help manufactories to transport important materials and reduce contact risk. The aim of this paper is to design a tracking control implementation system for the mobile robot, Turtlebot2. When it is passing a road crossing, the visually based SLAM system can recognize a moving target and predict its speed via a neural network-based speed estimation system. According to the robot velocity estimation result, the optimized passing routine and speed comes out. Similar to a human driving a car, the control system should drive the robot going through a crossing using the best passing program, which includes the following important elements, velocity, acceleration and desired path.