The speed and efficiency of Python programs is insufficient and the issue of their excessive memory usage and the running time problem appear in some cases. In the current study, the IPLS-XGBoost algorithm was adapted in optimisation framework complimented by the feature selection and dimension reduction, the gradient boosting tree was optimized, and the data pre-processing and caching, and parameter tuning strategies improved. Using these ways, we have extensively mitigated the Python applications. The trial results of the program after IPLS-XGBoost optimization suggest the program test data runs 0.8–1.5 s. The utilization of the integrated-problem-learning-strategy (IPLS-XGBoost) optimizer significantly improves the computational speed and performance of the Python software code based on the classification accuracy. Thanks to the feature selection and dimensionality reduction which decreased the complexity of the feature space, where the number of coefficients reduced, the speed of execution of the program was accelerated.