The purpose of this study was to estimate the differences in neuromuscular fatigue among prolonged pedalling exercises performed at different pedalling rates at a given exercise intensity. The integrated electromyogram (iEMG) slope defined by the changes in iEMG as a function of time during exercise was adopted as the measurement for estimating neuromuscular fatigue. The results of this experiment showed that the relationship between pedalling rate and the means of the iEMG slopes for eight subjects was a quadratic curve and the mean value at 70 rpm [1.56 (SD 0.65) microV.min-1] was significantly smaller (P < 0.01) than that at 50 and 60 rpm [2.25 (SD 0.54), and 2.22 (SD 0.68), respectively]. On the other hand, the mean value of oxygen consumption obtained simultaneously showed a tendency to increase linearly with the increase in pedalling rate, and the values at 70 and 80 rpm were significantly higher than those at 40 and 50 rpm. In conclusion, it was demonstrated that the degree of neuromuscular fatigue estimated by the iEMG changes for five periods of prolonged pedalling exercise at a given exercise intensity was different among the different pedalling rates, and that the pedalling rate at which minimal neuromuscular fatigue was obtained was not coincident with the rate at which the minimal oxygen consumption was obtained, but was coincident with the rate which most subjects preferred. These findings would suggest that the reason why most people prefer a relative higher pedalling rate, even though higher oxygen consumption is required, is closely related to the development of neuromuscular fatigue in the working muscles.