Traditional mechanical fault diagnosis methods often need to handle the collected fault wave signals， and then combine neural network to extract and classify the features. It is not only complex in process， time consuming， but also low in recognition accuracy. Therefore， this paper uses one-dimensional convolutional neural network （1D-CNN） to extract and classify the experimental vibration data of gear fault of an aeroengine， for establishing the 1D-CNN model of gear fault and diagnosing the bearing fault. From the test and analysis results， the accuracy of gear classification by using the neural network model is up to 80%， which is 15.07% higher than that of 63.9% of the traditional back propagation neural network， and the accuracy of this method is improved by 15.89% compared with the classification by support vector machine （SVM）. This method can directly use the wave vibration signal as input， and output the final classification results through a series of operations such as convolution and pooling. It simplifies the traditional tedious steps of signal processing and machine learning diagnosis， which provides a feasible method for aeroengine fault diagnosis.