Brown bear Ursus arctos population density was estimated for a 21,178-km² study area in southwest Alaska. Estimates were obtained using an aerial line-transect method that allows for peak detection to be both off the transect line and less than 100%. Data collection required five small aircraft with 2-person crews. Surveys were flown in 10-d windows to capture the period after den emergence but prior to full green-up. Surveys were flown in two consecutive years in order to detect sufficient bear groups to support the estimation. The study detected 197 bear groups (330 bears) in 969 aerial transects averaging 24.8 km long and with a strip width of 728 m. Estimated population density in the study area was 40.4 bears/1,000 km² (95% CI  =  31.4–54.5); estimated density of independent bears was 27.3 bears/1,000 km² (95% CI  =  21.4–34.4). Assuming similar estimate precision, repeating the survey could detect population changes of approximately 38% or larger with a power of 80%. We find the method described here suitable for regions of relatively high bear population densities or detection rates.