Drug-drug interaction (DDI) prediction poses a critical challenge in pharmacology and clinical practice, necessitating accurate and interpretable models. While existing computational approaches often concentrate on integrating diverse data sources and leveraging advanced machine learning techniques, they frequently lack interpretability, hindering adoption in clinical decision-making. To address this limitation, a novel framework is proposed, combining Graph Neural Networks (GNNs) with Explainable Artificial Intelligence (XAI) techniques for DDI prediction. This approach captures the complex relationships between drugs and associated entities within knowledge graphs (KGs) and provides transparent explanations for the model’s predictions, empowering clinicians with actionable insights. By integrating XAI techniques into the GNN-based framework, the interpretability of the predictive model is enhanced, enabling clinicians to understand the underlying reasons behind predicted DDIs and make informed treatment decisions. Experimental evaluations on diverse datasets demonstrate the efficacy and interpretability of the approach, highlighting its potential to advance patient conclusions and augment irrefutable decision-making in pharmacological practice. The Proposed Framework exhibits a peak accurateness of 0.82, trailed by MF with 0.76 and RW with 0.77. Similarly, in F1-score, recall, AUC-ROC, and precision metrics, the Proposed Framework outperforms MF and RW. These results demonstrate the superior performance of the Proposed Framework in both datasets, indicating its efficacy in DDI prediction.