Production scheduling plays a decisive role in supply chain management, directly influencing the operational efficiency and competitiveness of companies. This study explores the effectiveness of the LightGBM algorithm for production scheduling on non-identical parallel machines, comparing it to algorithms such as logistic regression, KNN, decision tree, and XGBoost. LightGBM was chosen for its speed of execution and its ability to handle large amounts of data. The results show that LightGBM outperforms the other models in terms of RMSE, MAE, explained variance score, and R² score for regression tasks, as well as in classification accuracy for certain features. Its superiority is attributed to its ability to efficiently handle data complexity while reducing computational complexity through its leaf tree growth technique. This study highlights LightGBM's potential for improving the efficiency of supply chain management systems and the challenges associated with computational scalability for large datasets. The results suggest that LightGBM is a robust and effective solution to optimize production scheduling, paving the way for future research in this field.