According to the AE (Acoustic Emission, AE) signal that caused by the aircraft structural crack, this paper proposed a health diagnosis approach based on ARMA model and SVM (Support Vector Machine, SVM). Firstly, this method apperceives the healthy information of a real aircraft stabilizer with real time by using AE detecting system. And then, deal with the information collected. Secondly, built an ARMA model and extracting parameter of ARMA model to set up characteristic vector. Finally, design and apply SVM monitor to diagnose the aircraft structure healthy state. Experiments show that the monitor has good performance to recognize and diagnose the fatigue crack of aircraft stabilizer, this approach is proposed to be effective. And it presents a new kind of approach to monitor the health state of aircraft structure and can also be used in other complex system structure health monitoring.