Abstract: As the core component of shearer transmission device, the gear has a high failure rate, which threatens the underground safety production, and it is of great significance to carry out the fault diagnosis of shearer gear transport. Based on this, a method of gear fault diagnosis based on BP neural network is designed. Taking the characteristic value from gear vibration signal as the input neurons of the gear network, and four fault states as the output neurons, the BP neural network model training. Comparing the diagnostic test results, the results show that this diagnostic method can accurately diagnose gear fault, with high accuracy, and can meet the needs of gear fault diagnosis, which is worth popularizing and applying.