The performance of helicopter surveys for estimating population densities of red kangaroos ( Macropus rufus ), eastern grey kangaroos ( Macropus giganteus ) and common wallaroos ( Macropus robustus ) was investigated by comparing line-transect density estimates obtained from helicopter surveys with those from ground (walked) surveys. Comparisons were made at four sites in central western Queensland (areas with high densities of red kangaroos and common wallaroos) that were surveyed during winter and summer from December 1991 until February 1995, and one site in south-eastern Queensland (an area of high eastern grey kangaroo density) surveyed annually during autumn from March 1991 until March 1994. Helicopter surveys generally recorded lower sample sizes than did ground surveys (means ± s.e = 34 ± 6%, 33 ± 9% and 76 ± 2% lower for red kangaroos, eastern grey kangaroos and wallaroos, respectively). Density estimates obtained from the helicopter surveys were not significantly different from those obtained from ground surveys for both red and eastern grey kangaroos as assessed by repeated-measures ANOVA and regression analysis. However, helicopter surveys of common wallaroos consistently returned density estimates about half those of ground surveys. The relationships between the two methods did not differ between winter and summer for any species. The conventional aerial survey method for kangaroos of strip transects from fixed-wing aircraft has limited ability to adjust for varying sightability conditions. Therefore, helicopter surveys with line-transect sampling are an attractive alternative.