We used a mechanistic movement model within a Bayesian framework to estimate survival, abundance, and rate of increase for a population of humpback whales ( Megaptera novaeangliae ) subject to a long-term photographic capture–recapture effort in southeastern Alaska, USA (SEAK). Multiple competing models were fitted that differed in movement, recapture rates, and observation error using deviance information criterion. The median annual survival probability in the selected model was 0.996 (95% central probability interval (CrI): 0.984, 0.999), which is among the highest reported for this species. Movement among areas was temporally dynamic, although whales exhibited high area fidelity (probability of returning to same area of ≥0.75) throughout the study. Median abundance was 1585 whales in 2008 (95% CrI: 1455, 1644). Incorporating an abundance estimate of 393 (95% confidence interval: 331, 455) whales from 1986, the median rate of increase was 5.1% (95% CrI: 4.4%, 5.9%). Although applied here to cetaceans in SEAK, the framework provides a flexible approach for estimating mortality and movement in populations that move among sampling areas.