Diesel engine is an extremely important equipment of the ship, and because of the complex structure of diesel engine, it is easy to produce faults and has a huge impact on the normal operation of the ship. In addition, some fault data of diesel engine are difficult to obtain, so the method of using fault simulation model to obtain fault data of diesel engine is proposed to support the training of fault diagnosis model. In this paper, the AVL BOOST simulation model is used to obtain the fault-related data set, and then the model is trained to realize and analyze the intelligent algorithms that can diagnose diesel engine faults, including random forest, SVM, and BP neural network. Experiments show that the fault diagnosis algorithm can fit the data well and diagnose the fault results, among which BP neural network has the best performance. This paper can provide a research direction for diesel engine fault diagnosis.