Seawater desalination using reverse osmosis (RO) technology provides cost-effective solutions for clean water supply. To reduce energy consumption, optimal RO network (RON) designs are generated by either heuristics approach or mathematical programming. This study develops a P-graph approach for RON synthesis for freshwater production. A superstructure based on predefined RON components (pumps, power recovery turbines, and RO units) is developed and optimal and sub-optimal network structures are determined by minimizing the network’s total annualized cost (TAC). Two case studies are investigated to demonstrate the capability of the P-graph approach. The first case study is a RON based on El-Halwagi (1992) which considers two pumps, two RO units, and two turbines. The second case study uses Evangelista’s (1985) RON configuration, which is a similar network type with only one pump causing the permeate to flow directly to the secondary RO unit. Results show that the optimal network for Case 1 has a TAC of 387,770 USD and an energy consumption of 28,861 MJ/d. The optimal network for Case 2 resulted in a TAC of 359,352 USD with an energy consumption of 24,937 MJ/d. The optimal structure of Case 2 removed the second stage of the system which decreased the TAC and energy consumption of the system. Through the P-graph approach, it can be observed that the framework serves as an alternative method for designing RONs with the advantage of being able to generate alternative process topologies for detailed engineering evaluation.