针对城市交通“智能运输系统” ,提出了基于改进BP神经网络理论模型的路面交通流动态时序的预测算法。在BP算法的自适应学习率 ,在动量法优化网络收敛性等方面 ,进行了深入研究 ,并改进了基本BP算法中的收敛速度慢和易陷入局部最小点等问题。文章给出了基于改进BP算法的交通流动态时序的预测算法仿真实验 ,结果验证了该算法的可行性和先进性。在交通流时序预测方面有一定的应用价值 In view of the “intelligent transport system” of urban transport, a prediction algorithm of dynamic timing of road traffic flow based on the improved BP neural network theory model is proposed. In the aspect of adaptive learning rate of BP algorithm and optimization of network convergence by momentum method, some problems such as slow convergence speed and easy falling into local minimum in basic BP algorithm are improved. The article gives a simulation experiment of traffic flow dynamic timing prediction algorithm based on improved BP algorithm. The results verify the feasibility and advancedness of the algorithm. It has certain application value in traffic flow forecasting