In order to solve these problems of traditional selecting aircraft spare parts method, such as more relying on human experience, low accuracy and low operational efficiency, a new method of selecting aircraft spare parts by using XGboost algorithm was proposed. The classification feature system was established. The XGboost algorithm was used to sort, analyze and screen the features, and the simplified version of the classification feature system was constructed. The k-fold cross validation method and empirical reference were used to group and train the sample data, and the results were compared with GBDT, RF, Adaboost and other classification algorithms. The results show that the XGboost algorithm can reduce the interference of human factors, and it is efficient, scientific and optimal.