In the continuously growing field of cybersecurity, Distributed Denial of Service (DDoS) attack detection is a continuous challenge. This study presents a hybrid machine learning framework that enhances detection accuracy and interpretability by integrating multiple machine learning models with explainable AI techniques. In our proposed model, Gradient Boosting and Extreme Gradient Boosting (XGBoost) shows excellent performance and SHapley Additive exPlanations (SHAP) to identify key features enabling a transparent understanding of the factors contributing to attack identification. The dataset undergoes meticulous preprocessing, including missing value dealing, encoding for categorical values, and numerical feature normalisation. The proposed Hybrid Ensemble model (GB+XGB) demonstrates strong performance, achieving a perfect precision score of 100% and accuracy of 99.96%. The comprehensive analysis demonstrates the efficacy of hybrid models in achieving high detection accuracy and transparency through explainable AI using SHAP analysis. This research significantly contributes to providing a robust, interpretable framework for DDoS attack detection, potentially informing future developments in threat detection systems and enhancing the security infrastructure.