In this study, the 5G long-term evolution (LTE) system and the Doppler migration technique are combined to investigate a non-line-of-sight collision avoidance warning system. The 5G LTE system is used as the communication networks of mobile cars and the Doppler migration technique is used to quickly detect the relative speeds and driving directions (coordinates) of target cars. Then, MATLAB software is used as simulation software and a straight expressway is used as the simulation scenario to simulate the non-line-of-sight collision avoidance warning system. A car is used as the observer and other cars with different conditions are used as tested objects in the simulation of the non-line-of-sight collision avoidance warning system based on the investigated 5G mobile communication network. The starting coordinates, starting speeds, accelerations, driving directions, and simulation time of all the cars are input in MATLAB software as the parameters to generate the moving coordinates, distances, speeds, and angles of all the cars in the simulation process. The simulation time is 15 s under different conditions, and the changes in both the Doppler shift and its slope between the observer car and other cars are observed and analyzed to judge the collision conditions. The analysis results show that the investigated system can generate an avoidance warning before the collision occurs.