Abstract: The chaotic characteristics of urban traffic are analyzed, and the error feedback coefficient is introduced to improve one-rank local-region method of chaotic time series prediction methods and the forecasting method based upon the largest Lyapunov exponents. Both of the two improved methods are successfully used in real-time prediction of traffic flow. The results show that the two forecasting methods can be used in the prediction of traffic flow with considerably high accuracy, and the improved forecasting method based on the largest Lyapunov exponents is more suitable for forecasting traffic flow.