Considering the continuous threat of terrorist attacks on vital structures, it is imperative to enhance their resilience to blast impacts. Current analytical approaches are costly and complex, necessitating a more streamlined method to evaluate structures under such threats. This research addresses this by introducing a machine learning (ML) model that predicts the highly nonlinear behaviour of reinforced concrete (RC) slabs under blast loadings. A database with 936 samples, including both experimental and numerical data, was carefully created for this study. The investigation scrutinized eight ML algorithms, refined them to four based on their performance, and optimized them using grid search, genetic algorithm, and particle swarm optimization (PSO). The gradient boosting-PSO hybrid model emerged to be superior, with a remarkable 91% accuracy in predicting maximum deflection. Further, a comprehensive influence analysis was conducted using the SHapley Additive exPlanations (SHAP) method to understand the contributions of various input parameters, pinpointing scaled distance and panel thickness as critical factors. This study, besides offering a rich database, also serves as an educational tool, shedding light on hyperparameter optimization techniques and SHAP analysis. The research promises a robust and interpretable ML model poised to significantly influence the practical engineering domain in improving the structural design of RC slabs facing blast impacts.