With the development of intelligent transportation, driverless is another important means that will utterly change traditional travels in the future. In order to adapt to the large-scale and complex traffic environment of driverless, we propose a dynamic bidirectional A* algorithm for driverless navigation planning. On the basis of this, theoretical simulations under different traffic flow conditions are realized, and the feasibility of the algorithm is verified. The vehicular self-organizing network is an important development direction of driverless in the future. In order to verify the performance of the algorithm in the vehicular self-organizing network environment, we adopt the OMNeT++ and SUMO bidirectional coupling, and perform joint simulation experiments on the open source framework Veins. In the traffic states with different traffic densities, the path planned by the dynamic bidirectional A* algorithm in the Vanet environment can reduce travel time and improve travel efficiency more effectively than that of the traditional bidirectional A* algorithm.