The purpose of this study was to test the reliability of a new handgrip exercise protocol measuring forearm oxygenation in 20 healthy subjects on two occasions. The retest took place 48 h later and at the same time of the day. The incremental exercise consisted of 2 min steps of cyclic handgrip contraction (1/2 Hz) separated by 1 min of recovery. The exercise started at 20% MVC, was increased with 10% MVC each step and was performed until exhaustion (69.5 and 73% MVC). Near infrared spectroscopy (NIRS) was used to measure deoxygenation (deoxy[Hb + Mb]) and oxygen saturation (SmO(2)) in the forearm muscles. Prior to the exercise protocol an arterial occlusion of the forearm was performed until deoxy(Hb + Mb) did no longer increase. Maximal increase in deoxy[Hb + Mb] during 10 s of each exercise bout was expressed relative to the occlusion amplitude. ICC was used to examine the test-retest reliability. Significant ICC's were reported at 50% (r = 0.466, p = 0.017) and 60% MVC (r = 0.553, p = 0.005). The group mean of the maximum increase in oxygen extraction was 45.6 ± 16.7% and at the retest 44.9 ± 17.0% with an ICC of r = 0.867 (p < 0.001) which could be classified (Landis and Koch 1979) as almost perfect. The absolute SmO(2) values showed reliable ICC's for every submaximal intensity except at 60% MVC. An ICC of r = 0.774 (p < 0.001) was found at maximal intensity. The results of the present study show that deoxy[Hb + Mb] and SmO(2) responses during this protocol are highly reliable and indicate that this protocol could be used to get insight into deoxygenation and oxygen saturation in a population with low exercise tolerance.