The purpose of this study was to analyze the test-retest reliability of the median frequency (MDF) and root mean square (RMS) used to determine the onset of neuromuscular fatigue (NF) during sustained fatiguing contractions of the quadriceps. Eighteen healthy men were tested on two days, and electromyographic (EMG) signals were recorded from the vastus medialis (VM), rectus femoris (RF) and vastus lateralis (VL) during sustained isometric contractions at 20 and 70% of maximum voluntary contractions (MVC) held until exhaustion. The reliability of endurance time was excellent at 20% MVC and poor at 70% MVC. EMG variables were evaluated: (1) at the beginning of the test; (2) at NF; and (3) at the end of the test. The NF time values presented poor reliability. The MDF has shown, in general, poor reliability at 20 and 70% MVC, whereas the RMS reliability presented better results for both loads, especially for RF, followed by the VM and VL muscles. The MDF and RMS values extracted from NF showed poor reliability at 20 and 70% MVC, which suggests caution in using these variables extracted from the EMG signal to determine the onset of NF.