Population parameters of American martens (Martes americana) are often monitored to guide management actions. In order to better understand changing marten population demographics, we estimated marten abundance and several population parameters on a portion of northeast Chichagof Island (NCI), Southeast Alaska, USA, using capture–mark–recapture methods in combination with radiotelemetry and compared these data with demographic estimates derived from the trapper catch from the larger NCI area. We found that capture probability remained relatively high across trapping sessions by using a standard live-trapping protocol, and we found no evidence of capture probability being consistently affected by population or environmental factors. Estimated marten numbers varied greatly over the period, ranging from a low of 14.1 martens (0.17/km²) during winter 1997–1998 to a high of 45.6 martens (0.55/km²) during winter 1995–1996. The annual trend was for decreasing numbers from winter 1990–1991 to winter 1992–1993, then increasing numbers through winter 1995–1996. By winter 1996–1997, numbers had dropped substantially and remained low through 1997–1998. All population parameters varied temporally, including annual survival (0.34–1.00), sex ratio (1.0–3.5), mean age (0.5–2.9), body condition (−0.08–1.35), and fecundity (0.44–2.70). Yearly trapper catches on NCI varied greatly from 19 to 354 martens (0.02 martens/km² to 0.31 martens/km²) because of changing trapper effort and marten abundance. We found mean ages of carcasses a good predictor of population mean age, but sex ratios were highly variable. Marten numbers and fecundity were strongly correlated with rodent abundance, especially long-tailed voles (Microtus longicaudus). By monitoring demographic parameters of the population or trapper-caught carcasses, managers can gain insight into temporal population dynamics. Also, total captures of individuals may provide a useful estimate of marten numbers without the expense of radiocollaring and tracking individuals because population estimates were highly correlated with number of individuals captured during a standard survey. Because of their high vulnerability to trapping, close monitoring of marten populations is important for the conservation and sustained-yield management of the species. A tracking strategy developed with input data on preseason abundance and age structure may be the best management approach, especially for small populations with limited immigration.