In view of the shortage of traditional traffic control and guidance model and algorithms, the traffic control and guidance model based on central coordination system (CCOS) is proposed. In this model, the short-term traffic prediction of past traffic data, the result of traffic incident detection and the real-time traffic flow data are used to design the traffic-oriented dynamic traffic information fusion. Moreover, using the neural network technology, the traffic control and guidance coordination model based on neural network system is presented. Its parameters are decided by the experiments. Finally, a number of typical local road networks are selected for simulation comparative experiments. The experiments show this model is feasible and effective.