Recently, deoxygenated hemoglobin (HHb) has been used as one of the most popular indicators of muscle O2 extraction during exercise in the field of exercise physiology. However, HHb may not sufficiently represent muscle O2 extraction, as total hemoglobin (tHb) is not stable during exercise. The purpose of this study was to measure various muscle oxygenation signals during cycle exercise and clarify which is the best indicator of muscle O2 extraction during exercise using NIRS. Ten healthy men performed 6-min cycle exercise at both moderate and heavy work rates. Oxygenated hemoglobin (O2-Hb), HHb, tHb, and muscle tissue oxygen saturation (SmO2) were measured with near-infrared spatial resolved spectroscopy from the vastus lateralis muscle. Skin blood flow (sBF) was also monitored at a site close to the NIRS probe. During moderate exercise, tHb, O2-Hb, and SmO2 displayed progressive increases until the end of exercise. In contrast, HHb remained stable during moderate work rate. sBF remained stable during moderate exercise but showed a progressive decrease at heavy work rate. These results provide evidence that HHb may not sufficiently represent muscle O2 extraction since tHb is not stable during exercise and HHb is insensitive to exercise-induced hyperaemia.