Historically, Lake Superior supported one of the largest and most diverse Lake Trout Salvelinus namaycush fisheries in the Laurentian Great Lakes, but Lake Trout stocks collapsed due to excessive fishery exploitation and predation by Sea Lampreys Petromyzon marinus. Lake Trout stocking, Sea Lamprey control, and fishery regulations, including a refuge encompassing Gull Island Shoal (Apostle Islands region), were used to enable recovery of Lake Trout stocks that used this historically important spawning shoal. Our objective was to determine whether future sustainability of Lake Trout stocks will depend on the presence of the Gull Island Shoal Refuge. We constructed a stochastic age-structured simulation model to assess the effect of maintaining the refuge as a harvest management tool versus removing the refuge. In general, median abundances of age-4, age-4 and older (age-4+), and age-8+ fish collapsed at lower instantaneous fishing mortality rates (F) when the refuge was removed than when the refuge was maintained. With the refuge in place, the F that resulted in collapse depended on the rate of movement into and out of the refuge. Too many fish stayed in the refuge when movement was low (0–2%), and too many fish became vulnerable to fishing when movement was high (≥22%); thus, the refuge was more effective at intermediate rates of movement (10–11%). With the refuge in place, extinction did not occur at any simulated level of F, whereas refuge removal led to extinction at all combinations of commercial F and recreational F. Our results indicate that the Lake Trout population would be sustained by the refuge at all simulated F-values, whereas removal of the refuge would risk population collapse at much lower F (0.700–0.744). Therefore, the Gull Island Shoal Refuge is needed to sustain the Lake Trout population in eastern Wisconsin waters of Lake Superior.