In 2003, a reintroduction program was initiated at Badlands National Park (BNP), South Dakota, USA, with swift foxes (Vulpes velox) translocated from Colorado and Wyoming, USA, as part of a restoration effort to recover declining swift fox populations throughout its historical range. Estimates of age-specific survival are necessary to evaluate the potential for population growth of reintroduced populations. We used 7 years (2003–2009) of capture–recapture data of 243 pups, 29 yearlings, and 69 adult swift foxes at BNP and the surrounding area to construct Cormack–Jolly–Seber model estimates of apparent survival within a capture–mark–recapture framework using Program MARK. The best model for estimating recapture probabilities included no differences among age classes, greater recapture probabilities during early years of the monitoring effort than later years, and variation among spring, winter, and summer. Our top ranked survival model indicated pup survival differed from that of yearlings and adults and varied by month and year. The apparent annual survival probability of pups (0.47, SE = 0.10) in our study area was greater than the apparent annual survival probability of yearlings and adults (0.27, SE = 0.08). Our results indicate low survival probabilities for a reintroduced population of swift foxes in the BNP and surrounding areas. Management of reintroduced populations and future reintroductions of swift foxes should consider the effects of relative low annual survival on population demography.