The restoration of lake trout Salvelinus namaycush in Lake Michigan has been an elusive goal of resource management agencies in the Great Lakes region. In this study, we estimated annual survival rates of adult lake trout from an area in northwestern Lake Michigan known as the Clay Banks refuge. We tagged and recaptured fish with gill nets during the fall spawning season (N = 12,175; 1983–1989 and 1991–1993) and with pound nets in the spring (N = 52,035; 1984–1990 and 1992–1993). We fit Cormack–Jolly–Seber models to the two sets of data. We had insufficient data to analyze annual differences in survival rates of fall-tagged fish, but we were able to estimate an overall annual survival rate of 0.67. Annual survival rates of spring-tagged fish varied between 0.53 and 0.88 and increased after 1987–1988. In addition to the mark–recapture studies, we analyzed catch rates of lake trout from gill-net and pound-net surveys to estimate survival rates using catch curve analyses; these annual rates were generally lower than those estimated from mark–recapture analyses of tagged fish. However, survival rates of lake trout from the Clay Banks refuge appeared to meet the minimum rate believed necessary for restoration of this species in Lake Michigan. Furthermore, adult survival rates have been increasing in recent years, and lake trout restoration in Lake Michigan is not hampered by low survival of adult fish. We hypothesize that the recent decrease in abundance of adult lake trout is primarily due to decreases in survival rates of lake trout younger than 6 years.