This work presents a comparison between centralized and decentralized control for cooperative distributed model predictive control scheme for linear systems. The presented scheme compensates the lack of accuracy and computational efficiency of model based technique for the geographically distributed controllers.The advantages of this control strategy is the ability to converges to the centralized optimal solution (Pareto optimum) and generate a control feedback that able to insure the closed-loop stability. A process of two subsystems is presented for simplicity, however, the strategy has the flexibility to be extended for any finite number of subsystems. © 2018 IEEE.