With the rapid development of aero-techniques, the conventional fault diagnosis and maintenance technique are not fit for the demand of modern civil aircraft. Data mining picking up connotative, unknown but latent valuable rules, is a strong tool for gaining knowledge. An inference engine based on the association data mining technique is studied and designed to solve the current conflict in reasoning matches and combinatorial explosion problem. The inference engine is used to analyze history maintenance record, gain knowledge the relationship between symptom and root cause hide behind the big data, sort the root causes according to its probability. The relationship between symptom and root cause, and the sort of root causes are helpful to fault diagnosis and decision-making.