Using BAM neural networks for aircraft fault diagnosis system in the center,provided by FTA support.FTA to a neural network to provide the necessary preprocessing,on the other hand for neural network training samples collection.Fault diagnosis model is shown in Figure 1,where FTA quantization module completes the fault event encoding,BAM neural network training samples; training sample database storage FTA provided training data samples; BAM module completes the fault knowledge base construction,as well as in the knowledge base based on the use of associative memory result of diagnosis; fault knowledge base memory for storing BAM neural network fault diagnosis correlation data; the interpreter then explains to the user the result of reasoning.