We examined whether transcranial magnetic stimulation (TMS) delivered to the motor cortex allows assessment of muscle relaxation rates in unfatigued and fatigued knee extensors (KE). We assessed the ability of this technique to measure time course of fatigue-induced changes in muscle relaxation rate and compared relaxation rate from resting twitches evoked by femoral nerve stimulation. Twelve healthy men performed maximal voluntary isometric contractions (MVC) twice before (PRE) and once at the end of a 2-min KE MVC and five more times within 8 min during recovery. Relative (intraclass correlation coefficient; ICC 2,1) and absolute (repeatability coefficient) reliability and variability (coefficient of variation) were assessed. Time course of fatigue-induced changes in muscle relaxation rate was tested with generalized estimating equations. In unfatigued KE, peak relaxation rate coefficient of variation and repeatability coefficient were similar for both techniques. Mean (95% CI) ICC 2,1 for peak relaxation rates were 0.933 (0.724–0.982) and 0.889 (0.603–0.968) for TMS and femoral nerve stimulation, respectively. TMS-induced normalized muscle relaxation rate was − 11.5 ± 2.5 s −1 at PRE, decreased to − 6.9 ± 1.2 s −1 (− 37 ± 17%, P < 0.001), and recovered by 2 min post-exercise. Normalized peak relaxation rate for resting twitch did not show a fatigue-induced change. During fatiguing KE exercise, the change in muscle relaxation rate as determined by the two techniques was different. TMS provides reliable values of muscle relaxation rates. Furthermore, it is sufficiently sensitive and more appropriate than the resting twitch evoked by femoral nerve stimulation to reveal fatigue-induced changes in KE.