In the process of fault diagnosis of analog circuits, there are some problems which are insufficient extraction of fault feature information and feature information redundancy, for this, in this paper, a fault diagnosis method for analog circuits based on maximum overlapping discrete wavelet packet transform (MODWPT) and local Fisher discriminant (LFDA) is proposed. In this method, firstly, MODWPT is used to original signal processing and fault features extraction of the analog circuit. Then, due to the redundant information in the high-dimensional feature set, which is not conducive to pattern classification and recognition, LFDA is used to reduce the dimension of the high-dimensional feature set and obtain a lower-dimensional feature set which is more conducive to fault pattern recognition. Finally, support vector machine (SVM) is used as fault pattern recognition classifier, and a fault diagnosis model is constructed based on SVM. Through analog circuit simulation experiments, the maximum fault diagnosis accuracy of the proposed method can reach 99.17%, which verifies the effectiveness of the proposed method.