The willingness of highly educated women to choose a mate is discussed. First, based on the "one-to-one matching"mode of a well-known marriage platform in China, the training data set and test data set are constructed. Secondly, based on the CART decision tree algorithm, an interpretable model of the female's willingness to choose a mate with high education is constructed, which outputs the common characteristics of the female's willingness to choose a mate and the characteristics of the most popular male group. Then, based on the XGBoost algorithm, an inexplicable model of women with high academic qualifications' willingness to choose a mate is constructed, and a set of intelligent recommendation strategies is proposed, which can provide personalized member recommendation results according to the needs of different female members. The experimental results on the test data set show that the F1-Score of CART decision tree model is 78.1%, and the F1-Score of XGBoost model is 85.2%. The actual test results show that the success rate of the intelligent recommendation strategy is significantly higher than the manual recommendation of the matchmaker.