Attempts to determine the age of faulting recorded by fault rocks have been made using K-Ar, fission-track, electron spin resonance, and luminescence dating methods. Here we report the unambiguous reduction (resetting) of optically stimulated luminescence (OSL) signals of quartz gouge after friction experiments and estimate the seismological and geological conditions required for OSL signal resetting in natural fault zones. In the experiments, we used quartz sand with a particle size of <150 μm and equivalent dose of 31.5 ± 16.6 Gy. Friction experiments were conducted with a rotary-shear, high-velocity friction apparatus at slip rates (V) of 200 μm s −1 to 1.3 m s −1 , normal stress of 1.0 MPa, and displacement of 10 m. In the experiments conducted under dry conditions, the OSL signal starts to decrease from V = 0.25 m s −1 and becomes near zero at V ≥ 0.65 m s −1. OSL signal resetting is also observed in the experiment sheared at 1.3 m s −1 under water-added conditions. At V = 0.25 and 0.40 m s −1 , partial resetting occurs, which is characterized by coexistence of particles with and without an OSL signal. OSL signal intensity shows a strong correlation with applied power density and frictional heat during high-velocity friction, and the signal exponentially decreases with increasing power density and temperature. The power density required for partial and complete OSL signal resetting is~0.17 and 0.6 MW m −2 , respectively. Assuming a co-seismic fault slip rate of 0.6 m s −1 , the depths required for partial and complete resetting are expected to be ≥11 and ≥42 m.