In this work, a computationally efficient solution for constraint management of square multi-input multi-output (MIMO) systems is presented. The solution, referred to as the Decoupled Reference Governor (DRG), maintains the highly-attractive computational features of scalar reference governors (SRG) compared to Vector Reference Governor (VRG) and Command Governor (CG). This work focuses on square MIMO systems that already achieve the desired tracking performance. The goal of DRG is to enforce output constraints and simultaneously ensure that the degradation to tracking performance is minimal. DRG is based on decoupling the input-output dynamics of the system so that every channel of the system can be viewed as an independent input-output relationship, followed by the deployment of a bank of scalar reference governors for each decoupled channel. We present a detailed set-theoretic analysis of DRG, which highlights its main characteristics. A quantitative comparison between DRG, SRG, and the VRG is also presented in order to illustrate the computational advantages of DRG. Finally, a distillation process is introduced as an example to illustrate the applicability of DRG.