The interaction of MIMO(multi-input multi-output) systems limits the use of many advanced design methods suitable for SISO(single input single output) systems. And most of the research on the control of MIMO systems only pays attention to eliminating the influence of interaction, but the problems of systems control are ignored, which weakens the system's robustness, the design process is complex, and the order of the system is higher after decoupling. Therefore, we propose a controller design method for MIMO systems based on the coefficient diagram method(CDM) to achieve control objectives relatively in the view of utility. The problem of decoupling of MIMO systems is transformed into parameter optimization problems. For this purpose, Firstly, the object function and two linear constraints are proposed, and the target function is optimized by the particle swarm optimization(PSO) in the frequency domain to design the compensator to achieve decoupling. Secondly, the CDM can design the controller structure and parameters flexibly according to the system's actual performance on the premise that it considers the stability, response characteristics, and robustness of the system. Therefore, the controller structure design and parameter tuning are based on the coefficient diagram method. Finally, the effectiveness of this method is verified by simulation examples.