To identify population‐scale patterns of movement, coastal cutthroat trout Oncorhynchus clarkii clarkii tagged and marked (35 radio‐tagged, 749 passive integrated transponder [PIT]‐tagged, and 3,025 fin‐clipped) were monitored from June 1999 to August 2000. The study watershed, located in western Oregon, was above a natural barrier to upstream movement. Emigration out of the watershed was estimated with a rotating fish trap. Approximately 70% of recaptured coastal cutthroat trout with PIT tags and 86% of those with radio tags moved predominantly at the channel‐unit scale (2–95 m); fewer tagged fish moved at the reach scale (66–734 m) and segment scale (229–3,479 m). In general, movement was greatest in April as spawning peaked and lowest in October, when discharge was at its lowest. Only 63 (<1% of tagged and marked fish) coastal cutthroat trout were captured in the fish trap. Trap efficiency was about 33%, and the expanded estimate of emigrants between February and June was 173 fish. These results suggest that unit‐scale movement is common throughout the year and that reach‐ and segment‐scale movements are important during the winter and spring. Although movement in headwater streams is most common at the channel‐unit scale, restoration of individual channel units of stream may not benefit the population at the watershed scale unless these activities are undertaken in the context of the greater whole. Individual coastal cutthroat trout move great distances, even within the small watersheds in the Oregon Coast Range, and although these movements may be infrequent, they may contribute substantially to recolonization after stochastic extirpation events (e.g., landslides and debris flows). Management strategies that focus on maintaining and restoring connectivity in a watershed represent an important step toward protecting the evolutionary capacity of stream salmonids.