Reduced reproductive success of hatchery fish spawning in the natural environment will reduce the ability of stocking programs to enhance wild populations. We used DNA-based parentage assignment to compare the reproductive success of wild fish and first-generation hatchery fish from a smolt stocking program that used broodstock from within the naturalized steelhead Oncorhynchus mykiss population in a Minnesota tributary to Lake Superior. The reproductive success of hatchery females was significantly lower than that of wild females (approximately 60%) in all three study years; however, the reproductive success of hatchery males was only significantly lower in one year. Higher reproductive success of wild fish was attributed to greater probability of success (i.e., having at least one offspring) and not differences in numbers of offspring among successful parents. Generalized linear models indicated that run timing was associated with probability of success although this did not explain differences between hatchery and wild fish. Despite runs that extended 7–9 weeks, most successful adults arrived in the first three weeks of the run (85–98% of all successful females, 98–100% of successful males). The early part of the run corresponded to periods of high flow, which likely increased access to quality spawning and rearing habitat higher upstream in the system. Relationships between fish length and reproductive success were inconsistent. Managers may minimize potential environmental and genetic contributors to reduced performance by hatchery fish, but continued reliance on hatchery supplementation may hinder achievement of the long-term goal of a fishery supported largely by naturally reproducing populations.