Abstract: According to important of safety protection and early warning of industrial control network safety production in coal mine，this paper analyzes three main areas of coal mine industrial control network security protection:the safety protection system on the well,the underground transmission and communication,the implementation of underground control and related safety protection strategies.In the network security protection of digital monitoring system,we try to introduce the idea of large data analysis,using the LM neural network prediction method,using the packet direction of network data,the packet interval,the data bit of the CAN protocol as collection of information,and establish the architecture and model program design of the network anomaly prediction system.The experimental results show that trough increasing the number of hidden layer neurons in the 3 layer neural network from 3 to 5 and 7,the model is stable and the error of prediction accuracy is close to 0.007%.It can classify the network anomaly of the monitoring system well to realize the prediction and early warning power.