The use of the acetylene rebreathing technique (ART) to estimate cardiac output (Q) at rest and during high-intensity exercise on a cycle ergometer has been validated. The reproducibility of the ART in estimating Qmax during graded exercise test (GXT) on a treadmill has not been reported. The attempt of the current study was to investigate the reliability of the ART in measuring Qmax during GXT in a small sample. Five male college students were involved in this study. All subjects underwent GXT until exhaustion twice on two test days separated by at least one week. A modified acetylene rebreathing procedure and the polar heart rate monitor were used to measure Q and heart rate (HR) at rest and during maximal exercise. To test reproducibility of ART, the repeat t-test and correlation procedures were utilized. There were no differences in Q, HR, and stroke volume (SV) either at resting or maximal exercising between test 1 and 2 (p > 0.05). No correlation between duplicate tests was found for resting Q, HR, or SV. However, Qmax or SVmax in test 1 were significantly correlated to that in test 2 (r = 0.92 or 0.94 respectively) (p < 0.05). The coefficients of determination for repeat measurements were 0.85 and 0.88 respectively. The results demonstrated that estimated Qmax and SVmax between duplicate tests were nearly identical. It suggested that with a small sample size, the Qmax measured by ART was highly reproducible. The acetylene rebreathing is a reliable technique that can be used to estimate maximal cardiac output during GXT on a treadmill.Keyword(s): measurement/evaluation