||Assessment of marine protected areas (MPAs) requires the ability to quantify the effects of proposed MPA size and placement, habitat distribution, larval dispersal, and fishing on the persistence of protected populations. Here we describe a model-based approach to assessment of the contribution of a network of marine protected areas to the persistence of populations with a sedentary adult phase and a dispersing larval phase. The model integrates the effects of a patchy spatial distribution of habitat, the spatial scale of larval dispersal, and the level of fishing outside of reserves into a calculation of the spatial distribution of equilibrium settlement. We use the amount of coastline predicted to have equilibrium settlement rates that saturate post-settlement habitat as a response variable for the assessment and comparison of MPA network designs. We apply this model to a set of recently proposed MPA networks for the central coast of California, USA. Results show that the area of habitat set aside is a good predictor of the area over which population levels will be high for short-distance dispersers. However, persistence of longer distance dispersers depends critically on the spatial distribution of habitat and reserves, ranging from not persistent anywhere to persistent over a greater area than that set aside in reserves. These results depend on the mechanisms of persistence, with self-replacement supporting short-distance dispersers and network effects supporting long-distance dispersers. Persistence also depends critically on fishery status outside the MPAs, as well as how fishing effort is redistributed after MPA implementation. This assessment method provides important benchmarks, as well as a transparent modeling approach, for improving initial MPA configurations that may result from less-comprehensive rule- or habitat-based methods of designing MPAs.