Abstract: Coal mine for the current trend of increasing network traffic monitoring, in order to improve and enhance the quality of the network QoS, we propose a chaotic time series prediction based on the network traffic.From the phase space reconstruction, with the mutual information and false near the point of law determined by the delay time and embedding dimension, with a small amount of data the maximum Lyapunov index method, which proved that the chaotic time series of network traffic characteristics, and the establishment of The corresponding model, successfully made its forecast.Simulation results show that the method has high accuracy.