Abstract Since urban traffic congestion has been a serious problem for several decades, a traffic signal control algorithm based on the model focusing on the dynamic behavior of links is developed in this paper. In the optimization process, the difference between the input capacity and output capacity of a link is optimized to make full advantages of the green time duration. Besides, considering the diverse traffic demand and capacity of the arterial streets and the side streets, we mainly consider the situation where they have a different cycle time. Model predictive control(MPC) is adopted to solve the optimization problem in a rolling horizon style to improve the performance of the control scheme. The efficiency of it is demonstrated by simulation.