A network monitoring device is designed to improve the ability of the power station, which belongs to the access layer of power dispatching data network, to make accurate self-judgements on network faults and abnormal behaviors. Based on mirrored data monitoring technique, the device uses Ping to detect the data link on-off state of other connect-to-network devices and Tcpdump to capture data packages. In addition, a task scheduler for multiple endpoints monitoring based on non-deprivation static scheduling is designed. With the monitoring results of source and destination IP addresses, communication time and amount of transported data, the monitoring device can make judgements on the existing of abnormal situations such as illegal communication, data channel interruption and DoS attack. Testing results have verified that the device is reliable and effective, and can make accurate judgements on network faults and abnormal behaviors, which helps to enhance the ability of the power station to identify and control the working condition of its own network. This work is supported by Natural Science Foundation of Xinjiang Uygur Autonomous Region (No. 2017D01C030). “Single-phase Fault Monitoring and Diagnosis Location for Low-voltage Distribution Network based on Big Data”