Populations of wild lake trout Salvelinus namaycush have been extirpated from nearly all their historical habitats across the Great Lakes. Efforts to restore self-sustaining lake trout populations in U.S. waters have emphasized the stocking of coded-wire-tagged juveniles from six hatchery strains (Seneca Lake, Lewis Lake, Green Lake, Apostle Islands, Isle Royale, and Marquette) into vacant habitats. Strain-specific stocking success has historically been based on estimates of the survival and catch rates of coded-wire-tagged adults returning to spawning sites. However, traditional marking methods and estimates of relative strain abundance provide no means of assessing strain fitness (i.e., the realized contributions to natural recruitment) except by assuming that young-of-the-year production is proportional to adult spawner abundance. We used microsatellite genetic data collected from six hatchery strains with likelihood-based individual assignment tests (IA) and mixed-stock analysis (MSA) to identify the strain composition of young of the year recruited each year. We show that strain classifications based on IA and MSA were concordant and that the accuracy of both methods varied based on strain composition. Analyses of young-of-the-year lake trout samples from Little Traverse Bay (Lake Michigan) and Six Fathom Bank (Lake Huron) revealed that strain contributions differed significantly from estimates of the strain composition of adults returning to spawning reefs. The Seneca Lake strain contributed the majority of juveniles produced on Six Fathom Bank and more young of the year than expected within Little Traverse Bay. Microsatellite markers provided a method for accurately classifying the lake trout hatchery strains used for restoration efforts in the Great Lakes and for assessment of strain-specific reproductive success.