We conducted a mark-recapture (MR) analysis from 8 years (2005–2012) of photo-identification data collected systematically to investigate demographic parameters of a community of bottlenose dolphins (Tursiops truncatus) inhabiting the Patos Lagoon Estuary and adjacent marine coast in southern Brazil. Under the most parsimonious model of Pollock's robust design, which disregarded the effects of temporary emigration, the estimate of annual apparent survival was higher for adult females (0.97, 95% CI: 0.91–0.99) than for adult males (0.88, 95% CI: 0.75–0.94) and juveniles (0.83, 95% CI: 0.64–0.93), which may explain an observed bias in sex ratio (1 male:2 females) of known adult dolphins in this community. An increase in abundance of marked individuals was observed during the first 6 years of sampling when the number of new recruits surpassed mortality, followed by a remarkable decrease in the last 2 years when an inverse ratio of recruits/deaths occurred. Yearly changes in abundance varied from -0.1 to 0.07. Total abundance estimates were highly precise (the highest coefficient of variation was 0.053) and did not exceed 88 individuals. Abundance estimates were similar to a previous MR study conducted in the same area almost a decade earlier, suggesting a relative stable dolphin community over the last 14 years. The apparent stability in abundance, however, should be viewed with caution since this community would need a substantial mortality of at least 10% before a decline in abundance is detected with a desirable statistical power of 90%.