In the field of aviation, the flight skills of trainee pilots are directly related to aviation safety and operational efficiency. Enhancing the flight skills of trainees through training has been one of our school"s objectives and is a core method for improving the safety of civil aviation flights. Based on the training data of flight trainees from a branch of our school, we introduced the Pearson correlation coefficient to evaluate the strength of the relationship between features and the target variable. Based on the correlation coefficients, key factors influencing flight training were identified, and a novel decision tree model based on Pearson correlation coefficients was established. By optimizing the model with various thresholds and tree depths, its performance in accuracy, precision, recall, and F1 score was enhanced. Comparative analysis against classical machine learning models such as Random Forest, MLP, Logistic Regression, Decision Tree, and enhanced Gradient Boosting Decision Tree showed superior performance of the new model. This study not only provides effective guidance for flight training but also offers theoretical support for the evaluation of flight training.