Pedalling cadence influences respiratory frequency (fR) during exercise, with group III/IV muscle afferents possibly mediating its effect. However, it is unclear how exercise intensity affects the link between cadence and fR. We aimed to test the hypothesis that the effect of cadence on fR is moderated by exercise intensity, with interest in the underlying mechanisms. Ten male cyclists performed a preliminary ramp incremental test and three sinusoidal experimental tests on separate visits. The experimental tests consisted of 16min of sinusoidal variations in cadence between 115 and 55rpm (sinusoidal period of 4min) performed during passive exercise (PE), moderate exercise (ME) and heavy exercise (HE). The amplitude (A) and phase lag (φ) of the dependent variables were calculated. During PE, fR changed in proportion to variations in cadence (r = 0.85, P < 0.001; A = 3.9 ± 1.4 breaths·min-1; φ = -5.3 ± 13.9 degrees). Conversely, the effect of cadence on fR was reduced during ME (r = 0.73, P < 0.001; A = 2.6 ± 1.3 breaths·min-1; φ = -25.4 ± 26.3 degrees) and even more reduced during HE (r = 0.26, P < 0.001; A = 1.8 ± 1.0 breaths·min-1; φ = -70.1 ± 44.5 degrees). No entrainment was found in any of the sinusoidal tests. The effect of pedalling cadence on fR is moderated by exercise intensity-it decreases with the increase in work rate-and seems to be mediated primarily by group III/IV muscle afferents, at least during passive exercise.