With the development of autonomous driving technology, human driving vehicles and automated driving vehicles will coexist for a period of time, forming a new type of mixed transportation. Connected and automated vehicles could bring new control methods for coordinated driving of vehicles. Mixed traffic efficiency optimization control methods are more based on macro-flow control or local optimization of small groups. The potential of CAV for traffic optimization has yet to be developed. In this paper, first the cyber physical description of mixed traffic is given, and the dynamics of the physical topology of mixed traffic is revealed. Subsequently, a CPS-based mixed traffic pinning control method considering dynamic physical topology from the perspective of cyber-physical is given. This method adjusts the motion state of CAV to indirectly influence, induce and control the human driving vehicles in traffic, thereby optimizing the entire transportation system. The simulation experiment based on MATLAB verifies the effectiveness of this method, and the simulation experiment based on SUMO shows that the pinning control method proposed in this paper is applicable in more realistic scenarios.