This study addresses the critical issue of credit default prediction in consumer lending, aiming to enhance risk mitigation and lending decision optimization. The research introduces an Ensemble Methods framework that combines LightGBM, XGBoost, and LocalEnsemble modules to improve accuracy and generalization. Each module contributes uniquely by leveraging distinct feature sets, enhancing diversity and robustness. The ensemble model is evaluated on the American Express dataset, demonstrating superior performance compared to other models, including deep learning and machine learning approaches. The study also includes feature importance analysis, highlighting the key features that contribute to credit default predictions. The proposed framework sets a new benchmark for credit default prediction, offering a comprehensive and adaptable solution.This study addresses the critical issue of credit default prediction in consumer lending, aiming to enhance risk mitigation and lending decision optimization. The research introduces an Ensemble Methods framework that combines LightGBM, XGBoost, and LocalEnsemble modules to improve accuracy and generalization. Each module contributes uniquely by leveraging distinct feature sets, enhancing diversity and robustness. The ensemble model is evaluated on the American Express dataset, demonstrating superior performance compared to other models, including deep learning and machine learning approaches. The study also includes feature importance analysis, highlighting the key features that contribute to credit default predictions. The proposed framework sets a new benchmark for credit default prediction, offering a comprehensive and adaptable solution.