Deep learning technology has been increasingly becoming a popular and intelligent tool in the field of mechanical fault identification. This paper presents an application of Deep Boltzmann Machines in fault diagnosis. Firstly， based on a rolling bearing fault test platform， the depth analysis of Deep Boltzmann machine in the field of rolling bearing fault diagnosis was carried out. Then， the scheme was applied to the fault diagnosis of 20 kg turbojet engine. Finally， the accuracy of BPNN and SVM-based models is compared with the proposed method. The experimental results show that the accuracy presented by DBM model is highly reliable and applicable in fault diagnosis.