Purpose: This study aims to advance the field of marine propulsion dynamics by investigating coupled torsional and transverse vibrations in shaft systems and proposing a novel approach for their mitigation. The focus is on incorporating Magnetorheological (MR) dampers controlled by an Adaptive Neuro-Fuzzy Inference System (ANFIS) to enhance adaptability and responsiveness. Methods: A sophisticated numerical model, based on a modified Jeffcott rotor model, is introduced to forecast the intricate dynamics of the shaft system. A meticulously formulated mathematical model for the MR damper is presented and validated through experimental investigations. The MR damper is subjected to sinusoidal displacement inputs at varying current levels, and the resulting force–displacement profiles are analyzed. The proposed numerical model for the shaft system is validated through experimental analysis, and the integration of MR dampers controlled by ANFIS is investigated. Investigations are conducted at different rotational speeds (rpm) to evaluate vibration reduction and stability. Results: The semi-active control system, comprising MR dampers and the ANFIS controller, has demonstrated superior performance compared to passive strategies. The reduction in torsional vibrations by a percentage reduction index of 21.01% and transverse vibrations by 28.36% underscores the potential of this innovative approach in dynamically controlling marine propulsion systems. The ANFIS controller's ability to dynamically adjust damping force has proven crucial in ensuring optimal performance. Conclusion: This study highlights the potential of MR dampers controlled by ANFIS in enhancing dynamic control for marine propulsion systems. Findings contribute valuable insights into addressing vibration-related challenges, emphasizing potential resilience and optimization in the maritime industry. The presented semi-active control system demonstrates superior performance, paving the way for innovative approaches in marine engineering.