The purpose of this study was to assess the validity of predicting the maximal oxygen uptake (VO2(max)) of sedentary men from sub-maximal VO2 values obtained during a perceptually regulated exercise test. Thirteen healthy, sedentary males aged 29-52 years completed five graded exercise tests on a cycle ergometer. The first and fifth test involved a graded exercise test to determine VO2(max). The two maximal graded exercise tests were separated by three sub-maximal graded exercise tests, perceptually regulated at 3-min RPE intensities of 9, 11, 13, 15, and 17 on the Borg ratings of perceived exertion (RPE) scale, in that order. After confirmation that individual linear regression models provided the most appropriate fit to the data, the regression lines for the perceptual ranges 9-17, 9-15, and 11-17 were extrapolated to RPE 20 to predict VO2(max). There were no significant differences between VO2(max) values from the graded exercise tests (mean 43.9 ml x kg(-1) x min(-1), s = 6.3) and predicted VO2(max) values for the perceptual ranges 9-17 (40.7 ml x kg(-1) x min(-1), s = 2.2) and RPE 11-17 (42.5 ml x kg(-1) x min(-1), s = 2.3) across the three trials. The predicted VO2(max) from the perceptual range 9-15 was significantly lower (P < 0.05) (37.7 ml x kg(-1) x min(-1), s = 2.3). The intra-class correlation coefficients between actual and predicted VO2(max) for RPE 9-17 and RPE 11-17 across trials ranged from 0.80 to 0.87. Limits of agreement analysis on actual and predicted VO2 values (bias +/- 1.96 x S(diff)) were 3.4 ml x kg(-1) x min(-1) (+/- 10.7), 2.4 ml x kg(-1) x min(-1) (+/- 9.9), and 3.7 ml x kg(-1) x min(-1) (+/- 12.8) (trials 1, 2, and 3, respectively) of RPE range 9-17. Results suggest that a sub-maximal, perceptually guided graded exercise test provides acceptable estimates of VO2(max) in young to middle-aged sedentary males.