In addressing the distinct influence mechanisms of torsional stiffness, torsional damping, and moment of inertia on the torsional vibration behavior of transmission systems, a human-simulated intelligent control strategy based on hybrid Taguchi genetic algorithm is proposed for the torsional vibration of magneto-rheological (MR) transmission system with variable stiffness, variable damping and variable inertia. Initially, the research conducts an in-depth analysis of the effects exerted by torsional stiffness, torsional damping, and moment of inertia on the transmission system dynamics. Subsequently, a comprehensive dynamics model for the MR transmission system with adaptable stiffness, damping, and inertia parameters is formulated, grounded on the conceptualization of an MR torsional vibration absorber. Drawing from the dynamic optimization outcomes facilitated by the hybrid Taguchi genetic algorithm, control parameters governing torsional stiffness, torsional damping, and moment of inertia are refined. This refinement process underpins the design of a human-simulated intelligent controller distinguished by its partitioned and multimodal characteristics. Notably, this controller framework is crafted to accommodate considerations across both temporal and spectral domains. Following the controller's design phase, rigorous simulation analyses are conducted to assess its efficacy. Results substantiate that the human-simulated intelligent control mechanism, underpinned by the hybrid Taguchi genetic algorithm, proficiently attenuates torsional vibrations across a broad spectrum of frequency bands. Furthermore, it is observed to markedly enhance the overall output characteristics of the MR transmission system.