In this paper, a unique dynamic decoupling control strategy, based on the active disturbance rejection control framework, is proposed for square multivariable systems. With the proposed method, it is shown that a largely unknown square multivariable system is readily decoupled by actively estimating and rejecting the effects of both the internal plant dynamics and external disturbances. By requiring as little information on plant model as possible, the intention is to make the new method practical. Simulation results obtained on two chemical process problems show good performance in the presence of significant unknown disturbances and unmodeled dynamics.