This work implements model-based controls for the simulation of reverse osmosis (RO) desalination process for both servo and regulatory problems. Pressure from high-pressure (HP) feed pump and recycle ratio (RR) are manipulated inputs to the RO process, while permeate flow rate, concentration and pH are considered as output. Control configuration for this 2 x3 process is selected using regular interaction analysis methods. Two control strategies are adopted, namely, multiloop internal model control-proportional integral controller (IMC-PI) under centralized scheme and multivariable PI controller with decoupler under decentralized scheme. The non-square model for synthesizing the model-based control purpose is formulated using the mass balance continuity equation and model parameters are estimated. In both the cases, satisfactory results are obtained by reducing the fluctuation in the controlled parameters in the closed loop control systems. The performances of the two controllers are compared using the integral of absolute error (IAE) performance index criteria. Results with MPC are encouraging for implementation on desalination process. (C) 2014 Elsevier B.V. All rights reserved.