Stock contributions to annual harvests provide key insights to conservation, especially in fish species that return to specific spawning sites and may establish genetically distinct populations. In this context, genetic stock identification (GSI) requires reference samples, yet sampling might be challenging as spawning sites could be in remote and/or unknown areas. Thus, any potential missing source population needs to be accounted for in management recommendations. Here, we (i) genotyped 1487 walleye (Sander vitreus) samples using a GT-seq panel of 336 single nucleotide polymorphisms and (ii) assessed individual migration distances from GPS records of fish harvested in two neighboring northern Quebec lakes (Mistassini and Mistasiniishish) important to the local Cree community. Samples were assigned to a source population using two methods, one requiring allele frequencies of known populations (RUBIAS) and the other without prior knowledge (STRUCTURE). Individual assignments to a known population reached 96% consistency between both methods. All five major source populations were identified in Mistassini Lake, but there was evidence of up to three small unsampled populations. Furthermore, Mistassini walleye populations were characterized by large differences in average migration distance with some remaining near their spawning rivers. In contrast, walleye in Mistasiniishish Lake were assigned with very high confidence to two populations with similar distribution throughout the lake. The complex population structure and migration patterns in the larger Mistassini Lake suggest a more heterogenous habitat and thus, greater potential for local adaptation. This study highlights the importance of combining analytical approaches to improve GSI studies for conservation practices.