According to the information system of electric power network is prone to traffic congestion problems, the need for traffic monitoring, improve the stability of power supply, this paper proposes a spectrum feature extraction and abnormal traffic filtering monitoring information system of electric power network data flow monitoring method based on the nonlinear characteristics of the network traffic data reorganization method, combined with adaptive spectrum detection method power spectrum analysis of information system of network traffic data to extract the spectral characteristics the amount of data, the blind separation of abnormal flow matched filter detector, realize the effective flow monitoring. The simulation results show that using the method of power information system network traffic monitoring data with real-time and accuracy, improve the stability of the power supply network.