In order to effectively diagnose aero-engine faults and ensure safety in aircraft flights, a method based on dynamic principal component analysis (PCA) and the improved support vector machine is put forth. It combines the advantages of dynamic PCA in the feature extraction and improves the support vector machine (SVM) in the fault diagnosis. The dynamic PCA method can complete the pre-treatment through de-noising, dimension reduction, and eliminating correlation on the processing variables. The improved SVM method can diagnose faults with the eigenvector. The proposed method can solve such problems as the lubrication system like, low accuracy of the aero-engine model, and limited measurement parameters, all of which are problems that can lead to low efficiency, ease of misdiagnosis, and other issues. A certain type of lubrication system of the aero-engine is taken as an example to verify the effectiveness of the proposed method. The results show that using the dynamic PCA and improved SVM fault diagnosis method can effectively improve accuracy and realize the fault diagnosis performance of the lubrication system of the aero-engine. Furthermore, it has good prospects for future application.