Purpose: This study aimed to evaluate the modulation and reliability of the vastus medialis (VM) and vastus lateralis (VL) H-reflexes in both passive and active conditions. Methods: Recruitment curves of VM and VL H-reflexes and M-waves at rest and during muscle contraction (30% of maximal voluntary contraction) were performed for 12 healthy males and were then repeated after 1 h, 1 day and 1 week. The maximal H-reflexes of each muscle were normalized to their respective maximal M-waves (H/M ratio) and absolute (CV) and relative (ICC) reliability were calculated. Results: The H-reflex was potentiated in active compared to passive condition and a higher H-reflex occurrence (12 vs. 10 subjects) and amplitude (≈+150%) was found for the VM compared to the VL in active condition. The intra- (ICChour = 0.97) and inter-day (ICCday = 0.92; ICCweek = 0.92) reliability was poor for the passive VM H/M ratio due to high within-subject variations (CVhour = 52.2%; CVday = 69.8%; CVweek = 60.9%) whereas for the active condition the reliability, especially intra-day, was good (ICChour = 0.93 and CVhour = 12%; ICCday = 0.86 and CVday = 14.5%; ICCweek = 0.79 and CVweek = 19.7%). Conclusions: This study showed differential modulation of the H-reflex between vasti muscles of the quadriceps and a higher occurrence and reliability for the active VM H-reflex. One can therefore conclude that it seems more appropriate to evoke quadriceps VM H-reflex (rather than VL) during an isometric muscle contraction.