Abstract: Prediction of the traffic mutation in network for mine monitoring can enhance quality of service (QoS) and reduce network congestion. In view of the characteristics of traffic in mine monitoring network, data acquisition (SCADA) information is selected as indicator of chaotic signal and Lyapunov exponent method is used to verify the chaotic characters of the indicator. Duffing oscillator is employed to calculate the amplitude of the critical threshold of the oscillator ad and a predicting model is constructed. The error between simulation results and real data is between ±0.036 632, validating the accuracy and reliability of the predicting method.