For the characteristics of urban multi-intersection systems, such as strong nonlinearity, time-varying and periodicity, at the same time considering factors such as the inherent capacity of the road and the urgency of alleviating road congestion, under a distributed control architecture, a model free adaptive iterative learning signal timing scheme based on decentralized estimation and decentralized control is designed. By adjusting the signal timing of each intersection in real time, the traffic processing of each intersection is adjusted, and the traffic processing capacity of each intersection is balanced. Furthermore, it can improve the utilization rate of road resources and alleviate urban traffic congestion. Finally, the effectiveness of the proposed scheme is further verified through simulation analysis.