Abstract: The ocean reverberation has time variability and non stationary. The sea reverberation will have a destructive interference to the underwater communication network. Communication network fault and communication channel imbalance, the need for the underwater ship communication network fault data for effective identification, and improve the stability and reliability of the submarine communication. In the traditional method, the fault data is identified by using the matched filtering method. The performance is not good with the nonlinear variation of the ocean reverberation. A new method of fault data recognition based on sub-pre whitening matching detection of underwater ship communication network is proposed. The model of underwater ship communication network model and communication channel model is built. The fault data signal is decomposed into multiple narrowband signals. The signal is detected by using AR model. The results are compared with the data of different time. The method is used to detect the different time of the data. Simulation results show that the proposed algorithm can effectively improve the accuracy of fault data, and can improve the accuracy of the fault data in the communication network, and improve the quality of the underwater communication.