Abstract: Studies on the instability fault pattern recognition method based on acoustic emission technology in Q345B steel were under taken. The method is based on the acoustic emission counting rate which is as the data object. It uses the mathematical statistics to analyze the law of the change of the count rate data and advances that the second and third stages of the fatigue crack propagation process are identified by the discrete coefficients of the acoustic emission count rate. It also provides the relationship between the discrete coefficient of the Q345B steel acoustic emission rate and the amplitude of the stress intensity factor. The relationship can provide the basis for the identification of Q345B steel instability fracture pattern by using acoustic emission.