Abstract: At present, coal mine full scene monitoring system mainly depends on cloud computing for data processing, storage and decision-making. Cloud computing needs to process massive amounts of monitoring information in real time, which seriously affects timeliness and accuracy of system decision-making layer. In view of the above problem, a coal mine full scene monitoring system based on fog computing was proposed. Fog computing neural network is designed with neuron sensing nodes as a unit to alleviate the pressure of cloud computing data processing. In view of problem of premature convergence and local optimal solution of the node deployment method based on particle swarm optimization algorithm, improved particle swarm optimization algorithm is used to optimize the deployment of neuron sensing node to achieve network structure optimization. Simulation results show that compared with the classic PSO algorithm, the improved PSO algorithm can find the optimal solution faster, and the optimal value, the worst value, and the average value of overall communication coverage have increased by 3.19%, 3.31%, and 3.25%, respectively, which has the advantages of fast and effective convergence, strong adaptability and high stability.