Abstract: In order to solve the problems of low accuracy of monitoring data, weak data analysis capabilities, and low intelligenceof existing mining equipment, an intelligent safety monitoring and predictive maintenance system for mining equipment is proposed.The system preprocesses mining equipment safety monitoring data through an intelligent data pre-selection model based on the simulated annealing algorithm, filters out the data within the threshold range so as to eliminate outliers in the data. The intelligent safety predictive maintenance model based on convolutional neural network is used to perform dual-channel convolution using convolutional kernels of different sizes to diversify the extracted filtered data characteristics. By comparing them with the mining equipment historical status sample data, it is able to judge the mining equipment operation status and then provide corresponding equipment predictive maintenance solutions.The experimental results show that the system has high accuracy in judging the operation status and predictive maintenance of mining equipment.