<p>This paper contributes to the evolutionary fuzzy fusion model predictive control (EFFMPC) for real-time appearance-based mapping (RTAB-Map) in visual simultaneous localization and mapping (VSLAM) of autonomous mobile robots. An evolutionary fuzzy logic system is incorporated into the MPC framework to establish an optimized control strategy through hyperparameter tuning. This approach is particularly beneficial for controlling complex, nonlinear systems, especially those with uncertainties or where precise mathematical models are difficult to obtain. In this study, the proposed EFFMPC is employed to develop an intelligent RTAB-Map VSLAM system for autonomous mobile robots. A metaheuristic spotted hyena optimizer (SHO) is applied to systematically explore the high-dimensional hyperparameter space of fuzzy logic system and RTAB-Map, aiming to improve mapping consistency, localization accuracy and control performance. This EFFMPC technology is utilized in robotic SLAM research for autonomous mobile robots equipped with physical visual sensors, enabling closed-loop visual navigation and demonstrating superior robustness in unknown dynamic environments. The developed approaches are deployed on a differential-drive autonomous mobile robot with onboard vision sensors. Empirical evaluations, including benchmark comparisons against conventional SLAM configurations, demonstrate the efficacy of the EFFMPC and SHO-based optimization strategy in improving overall SLAM performance for real-world robotic applications.</p>