Managing spatially structured populations of imperiled species presents manychallenges. Spatial structure can make it difficult to predict populationresponses to potential recovery activities, and learning through experimenta-tion may not be advised if it could harm threatened populations. Adaptivemanagement provides an appealing framework when experimentation is con-sidered too risky or time consuming; we used such an approach for imperiledflatwoods salamanders at a Florida wildlife refuge. We represented thismetapopulation with category count models and used stochastic dynamic pro-gramming to identify optimal decision policies that weighed trade-offs betweenmetapopulation persistence and management costs. We defined possible wet-land categories in terms of habitat suitability and occupancy, specifiedcategory-specific management actions, and identified transition probabilitiesvia expert elicitation for two management strategies:“future”status quo (FSQ;frequent growing-season burns) and extra management actions (EMA; restora-tion, translocation, head-starting). We simulated metapopulation dynamicsusing the resulting optimal management policy and found that under model FSQ, occupancy steadily declined over time, indicating that populations wouldrapidly become extirpated; with model EMA, occupancy remained stable,suggesting that populations would persist only if additional actions are appliedand are effective. This approach can be used to identify optimal solutions whileaccounting for uncertainty and considering both habitat and populationdynamics, and to help managers make conservation decisions for populations at imminent risk of extinction.