Abstract: An identification method of the offshore platform tubular joints fatigue crack propagation was proposed. Wavelet allocation scalogram and wavelet energy coefficient method were applied for the feature extraction of acoustic emission(AE) signal of marine structures typical fatigue crack propagation and fracture at the same time. The method could improve the aggregation of scalogram by excluding noise. By making the typical tubular joints model fracture experiments of the offshore platforms and using PXWAE AE detector to monitor the whole process of the specimen fracture, the paper successfully extracted the AE signals of crack propagation and fracture process. The analysis results demonstrated that the variation of wavelet energy coefficient and the change of crack propagation state were in a good consistency and could characterize the AE signals characteristics of different stages of crack propagation accurately, intuitively, and clearly. It could be used effectively in online real-time monitoring and fault diagnosis of offshore platform tubular joints fatigue crack.