Poor blast fragmentation has direct effects not only on blasting costs but also on the efficiency of downstream processes. Toronto mine has recently encountered significant challenges in optimizing its blasting and fragmentation processes with an average fragmentation efficiency of 60%. The main goal of this research was to optimize blast fragmentation efficiency from an average of 60% to above 80%. The methodological approach proposes the development of a hybrid Artificial Intelligence model, Adaptive Neuro-Fuzzy Inference System (ANFIS). The proposed ANFIS model performance has been compared with Artificial Neural Network (ANN) and Multivariate Regression (MVRA) models after training. The model performance in prediction was tested using 5 statistical indices, Coefficient of Regression, Mean Square Error, Root Mean Square Error, Mean Absolute Error and Value Accounted For. Based on statistical indices, it was found that the ANFIS model performs better than the ANN and the MVRA models, achieving a coefficient of determination of 0.9955 compared to 0.9249 (ANN) and 0.9005 (MVRA). Additionally, Root Mean Square Error improved from 0.8305 (ANN) and 2.0738 (MVRA) to 0.2087 for ANFIS, demonstrating significant prediction accuracy gains. The sensitivity analysis shows that the burden and uniaxial compressive strength have the most effect on the blast fragmentation efficiency. The ANFIS model optimized by Particle Swarm Optimization algorithm (PSO) provided optimized input parameters to achieve a blast fragmentation efficiency of 84%. These optimized input parameters were used in the trial blast. WipFrag software was used to analyze and evaluate blast fragmentation performance based on images of blasted muck piles from the trial blast. From the WipFrag analysis a targeted particle size distribution of P80 at 15 cm by 15 cm was achieved and this demonstrated that the proposed methodology offers a practical framework for achieving the objectives. Finally, the results of this research provide strong evidence for the ongoing use of Artificial Intelligence and metaheuristics algorithms in mining operations to maximize efficiency, lower operating costs, and enhance general mine's profitability.