Abstract: The accurate real-time forecast of traffic volume is the premise and basement of the dynamic traffic control and guidance.In order to forecast the traffic volume more accurately,this paper establishes an optimization model of genetic neural network.It not only utilizes the advantages of the genetic algorithm which are global search and rapid convergence,but also makes use of the merits of the neural network which are nonlinearly describing,self learning and self adapting.It has been applied on a real road to forecast the traffic volume and the computer simulations have proved that the optimization model has great precision and feasibility.