Abstract: Planetary gearbox is an important transmission component of mechanical equipment, and its operation directly affects the operating status of the entire equipment. The convolutional neural network model is improved by introducing the batch normalization layer and the discarding layer, and a gearbox fault diagnosis model based on the improved convolutional neural network is proposed. Set up a gearbox experiment platform, and use the model to identify faults in the gearbox vibration signal. The experimental results show it can be known that the improved model can effectively identify and classify the different types of gearbox faults. The classification accuracy rate reached 99.2%.