Abstract: In order to achieve the purpose of extracting the characteristics of different states of gears effectively and classifying them reasonably, in this paper, a fault diagnosis method of planetary gearbox based on genetic algorithm to optimize the distance compensation factor is proposed. Firstly, the time domain discrete signal of the gear is mapped to the graph domain to obtain the graph signal. Then, the genetic algorithm is used to optimize the distance compensation factor, and the optimized distance compensation factor is used to correct the Mahalanobis distance, Finally, the K-median clustering algorithm is applied to evaluate and classify the feature index set, so as to achieve the purpose of accurate classification of gears in different states. The experimental results show that the graph signal reconstructed by the distance compensation factor optimized by genetic algorithm has better data structure. Compared with the feature extraction method in time domain, the feature index extracted by this method is more representative and the classification accuracy is higher.