To solve data imbalance and insufficient prior knowledge problems in using quick access recorder （QAR） for health management and abnormal detection of large passenger aircraft air-conditioning system， this paper carries out an algorithm based on improved fuzzy C-means （FCM） to evaluate components degradation status. This algorithm is to calculate the distance between fault state and normal state. Based on the physical characteristics of large passenger aircraft’s air-conditioning system， the distance function of FCM algorithm is improved， and the difference of the left and the right packs is adopted as the evaluation standard. The overfitting problem of current methods is effectively solved， and the improved FCM algorithm is more sensitive to the degradation of the components in the early stage. It can also effectively reflect the intermediate process of performance degradation. This study provids strong technical support for airlines in arranging maintenance plan and reducing operation cost.