Conserving species requires managing threats, including habitat loss. One approach to managing habitat loss is to identify and protect habitat in networks of reserves. Reserve network design is a type of resource allocation problem: how can we choose the most effective reserve network design given available resources? We undertook development and implementation of a patch dynamics model to allow us to evaluate proposed reserve networks in terms of ability to sustain populations of several taxa that are dependent on native prairie in the South Puget Sound region of Washington, USA. We used expert input to build a patch dynamics model for each taxon and used the model to examine probability of persistence in 50 years under a variety of reserve network designs, including the existing reserve network. Results suggest that the existing reserve network offers varying levels of protection for the different taxa, from desirable (>90% certain that the probability of persistence is ≥75% in 50 years) to negligible. We identified a reserve network that was >90% certain to protect all 6 taxa of interest, which would require a combination of land protection and translocations of taxa to new or existing reserves. Post hoc, we also identified possible hybrid alternatives, involving addition of new reserves and growth of existing reserves, that protected all 6 taxa without translocations. The approach we demonstrate is technically tractable and allows for the evaluation of any proposed reserve network design, thereby allowing a decision maker to evaluate a set of reserve networks that meet resource constraints and determine which of those best meets conservation objectives.