The traditional single-channel vibration signal diagnosis method could only collect partial information for local diagnosis, the multi-channel signal fusion was also very difficult due to the uncertain weight and poor real-time property, and the fault diagnosis method of diesel engine was hence put forward based on one-dimensional deep convolutional neural network（1DCNN） and dual-channel information fusion. By building a diesel engine fault-preset test bench, the dual-channel fault signals during the operation were collected by the sensors of engine different positions and 14 features such as the maximum value, minimum value, peak-to-peak value, mean value, rectified average value, variance, standard deviation and kurtosis were extracted respectively. A feature set matrix was built, the feature fusion was then conducted by using the principal component analysis (PCA), the 1DCNN was further used, and the diagnosis of different engine fault states was realized in the end. The experimental results show that the method can effectively identify different engine fault states. Compared with singlechannel signal diagnosis, the proposed dualchannel information fusion method has better effect in engine fault diagnosis.