Predicting Breast Cancer (BC) accurately and early is essential to improve patient outcomes. The hybrid exPlanations Artificial Intelligence (XAI) model for breast cancer detection and diagnosis was presented in this work. The hybrid model uses the Wisconsin Diagnostic Breast Cancer (WDBC) dataset to evaluate the effectiveness of Random Forest (RF) classifiers for binary classification of benign and malignant tumors. Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are used as feature selection methods before classification in order to increase model performance and generalization. The RF classifier with PSO-selected features achieved the highest accuracy of 98.25%, outperforming both the baseline RF (96.49%) and GA + RF (95.61%) models. These results demonstrate that PSO-based feature selection enhances the predictive capability of ensemble learning models in medical diagnosis. We also compare baseline RF with GA + RF and PSO + RF models utilizing standard measures such as precision, recall, accuracy, F1-score, and AUC. SHapley Additive exPlanations (SHAP) serves the purpose of interpretability, and ROC curves and confusion matrices validate performance. The PSO + RF model achieves the highest accuracy (98.25%) and AUC (0.984), indicating superior predictive capability.