American horseshoe crabs (Limulus polyphemus) have varied migration patterns and harvesting pressure throughout their range, potentially leading to regional differences in population dynamics. Here, a multi-state mark–recapture model was used to estimate annual survival and exchange rates of adult horseshoe crabs across three geographic regions in Long Island, NY (South Shore, North Shore, and Jamaica Bay areas). Under the New York Horseshoe Crab Monitoring program, a total of 22,525 adult horseshoe crabs were tagged and 879 (3.9%) unique recaptures were observed from 2007 to 2016. Model-averaged annual survival in the North Shore population was higher at 68% (95% confidence interval (CI) 61.9–73.4) when compared to the South Shore (56.8%, 95% CI 51.1–62.2) and Jamaica Bay (54.5%, 95% CI 47.0–61.7) regions. Differences in survival between the North Shore and South Shore may reflect the greater harvest pressure directed along the South Shore. Contrary to expectations for a primarily closed region, Jamaica Bay survival was low, but not attributable to reported harvest related activities. Annual movement from the Jamaica Bay into the adjacent South Shore region was 19.8% (95% CI 13.1–28.9), but annual exchange rates ranging from 0.5 to 5.0% were observed between other regions. For example, movement from the South Shore and North Shore into Jamaica Bay was 3.5% (95% CI 2.3–5.9) and 0.5% (95% CI 0.0–1.0), respectively. There was strong support for sex-specific differences in survival, primarily driven by the low survival of females in Jamaica Bay (33.8%, 95% CI 21.1–50.5). Our findings reveal potential management implications, such as regional survival differences within a uniformly managed stock, and net emigration from a predominantly closed to open harvest region reducing the effectiveness of a protected area.