Pressure retarded osmosis has the potential to produce renewable energy from natural salinity gradients. This work presents the fabrication of thin-film composite membranes customized for high performance in pressure retarded osmosis. We also present the development of a theoretical model to predict the water flux in pressure retarded osmosis, from which we can predict the power density that can be achieved by a membrane. The model is the first to incorporate external concentration polarization, a performance limiting phenomenon that becomes significant for high-performance membranes. The fabricated membranes consist of a selective polyamide layer formed by interfacial polymerization on top of a polysulfone support layer made by phase separation. The highly porous support layer (structural parameter S = 349 μm), which minimizes internal concentration polarization, allows the transport properties of the active layer to be customized to enhance PRO performance. It is shown that a hand-cast membrane that balances permeability and selectivity (A = 5.81 L m-2 h-1 bar-1, B = 0.88 L m-2 h-1) is projected to achieve the highest potential peak power density of 10.0 W/m2 for a river water feed solution and seawater draw solution. The outstanding performance of this membrane is attributed to the high water permeability of the active layer, coupled with a moderate salt permeability and the ability of the support layer to suppress the undesirable accumulation of leaked salt in the porous support. Membranes with greater selectivity (i.e., lower salt permeability, B = 0.16 L m-2 h-1) suffered from a lower water permeability (A = 1.74 L m-2 h-1 bar-1) and would yield a lower peak power density of 6.1 W/m2, while membranes with a higher permeability and lower selectivity (A = 7.55 L m-2 h-1 bar-1, B = 5.45 L m-2 h-1) performed poorly due to severe reverse salt permeation, resulting in a similar projected peak power density of 6.1 W/m2. © 2011 American Chemical Society.