Incorporating connectivity into the design of marine protected areas (MPAs) has met with conceptual, theoretical, and practical challenges, which include: 1) the need to consider connectivity for multiple species with different dispersal abilities, and 2) the role played by variable habitat quality in determining the spatial patterns of connectivity. We propose an innovative approach, combining biophysical modeling with a routinely-used tool for marine-reserve design (Marxan), to address both challenges by using ecologically-informed connectivity parameters. We showed how functional demographic connectivity for four candidate reef-associated species with varying dispersal abilities and a suite of connectivity metrics weighted by habitat quality can be used to set conservation objectives and inform MPA placement. Overall, the strength of dispersal barriers varied across modeled species and, also across species, we found a lack of spatial concordance of reefs that were high-quality sources, self-persistent, and stepping-stones. Including spatially-heterogeneous habitat quality made a considerable difference to connectivity patterns, significantly reducing the potential reproductive output from many reefs. We also found that caution is needed in combining connectivity data from modeled species into multi-species matrices, which do not perform reliably as surrogates for all connectivity metrics of individual species. We then showed that restricting the habitat available for conservation has an inequitable impact on different connectivity objectives and species, with greatest impact on betweenness centrality and long-distance dispersers. We used Brazilian coral reefs as a case study but our approach is applicable to both marine and terrestrial conservation planning, and offers a holistic way to design functionally-connected reserves to tackle the complex issues relevant to planning for persistence.