In order to research the failure of the recoil system of the artillery under extreme working situations, taking the recoil system as the research object, analyzing common failures during recoil process and establishing a dynamic motion model, a state recognition method based on time-domain envelope analysis is proposed. The algorithm uses wavelet transform to process the recoil data, and recognizes the working status of the recoil system through time-domain envelope analysis and real-time threshold monitoring methods, and finally establishes databases to store recoil data. In the two failure modes, the recoil displacement and speed far exceed the specified threshold range, which can effectively identify the working status of the recoil system, and the identification efficiency and accuracy are greatly improved.