This study examined inter-session and intra-session transcranial magnetic stimulation (TMS) reliability at two test stimulus intensities in the knee extensors. Strong and weak TMS was delivered via single- and paired- (3-ms and 100-ms inter-stimulus interval) pulses on the same day and different days. All stimuli were delivered during isometric contractions of the knee extensors at 20% of maximal voluntary force. Motor-evoked potentials (MEP) were assessed in quadriceps femoris muscles. Relative (intra-class correlation coefficient, ICC) and absolute (standard error of measurement, SEM) reliability and variability (coefficient of variation) were assessed. MEPs elicited by strong and weak single-pulse TMS had excellent relative reliability in all muscles as did weak short-interval and strong long-interval paired-pulse TMS (all ICC > 0.75). Conversely, relative reliability of strong short-interval and weak long-interval paired-pulse TMS was lower (ICC: 0.34–0.83 and 0.22–0.97, respectively). MEP size variability was lower (P < 0.05) and SEM comparable or lower in strong compared to weak TMS conditions. These results suggest single- and paired-pulse TMS at both strong and weak intensities are generally reliable in the knee extensors. Strong (or both strong and weak) single-pulse TMS is recommended. The results indicate using weak test pulses for short-interval and strong test pulses for long-interval paired-pulse TMS are recommended.