In an intelligent networked environment, vehicles can communicate and execute precise control schemes to pass through unsignalized intersections. In the previous work, we modeled the Petri net model of the intersection without conflict and deadlock. This model mainly ensures the safety of vehicles, making vehicles pass the intersection safely without conflict, but it does not consider efficiency. With the Petri net model, an improved genetic algorithm with priority is proposed to optimally sequence the vehicle passing an intersection in this paper. Experimental results show that the proposed method effectively balances safety and efficiency, that is, avoids traffic conflicts at intersections, and improves the efficiency of traffic flow.