The design of reverse osmosis (RO) network for desalination for a wide range of salinity and seawater temperature is considered here. We start with a flexible superstructure that contains all possible alternatives of a potential RO network and refine the network during the synthesis using model based optimization technique. The synthesis problem is posed to find out an optimal RO network which will minimize the total annualized cost consisting of capital costs (costs of sea water pre-treatment, pump, turbine and membrane modules) and operating costs (pumping energy, chemicals, maintenance and membrane replacement) while fulfilling a given fresh water demand and quality. The problem results in a mixed-integer non-linear programming (MINLP) problem due to the presence of both integer and non-integer decision variables. Outer approximation algorithm within gPROMS software is used to solve the optimization problem. The results revealed that the feed salinity and temperature have significant impact on the RO network design and operation. (C) 2012 Elsevier B.V. All rights reserved.