Conventional multivariable controls may face challenges in industrial implementation due to their computation intensity, controller complexity, and/or poor robustness. To this end, this paper developed a practical multivariable control method, consisting of inverted decoupling and decentralized active disturbance rejection controller (ADRC). Strong robustness is achieved with negligible computation and simple forms of the decoupler and controller. Moreover, the disturbance rejection is markedly accelerated. On the basis of the two-input-two-output (TITO) system description, first discussed are the practical advantages of inverted decoupling that can be easily extended to high dimensional systems. Particularly, a compensation method is proposed to make the inverted decoupling applicable for the processes with right-half plane zeros. Then, the ADRC is bridged to the PI controller and the internal stability and robustness are analyzed. The feasibility of implementing ADRC in industrial distributed control system (DCS) is verified experimentally. Moreover, the qualitative tuning rules are discussed and packaged as an interactive tool. Also addressed are the compatibility and complementarity of the combination of ADRC and inverted decoupling. Finally, simulation and experimental results demonstrate the efficacy of the proposed method.