The colorimetric patterns of color changes of skin bruises were examined in adult male Japanese volunteers wounded by a sharp pinch on the inside of the upper arm and forearm. In most cases the non-traumatic area near the bruise could be used for the control skin. By using a colorimeter, tristimulus and spectrophotometric methods were available to objectively analyze the skin discoloration. On the L* a* b* color space analysis, Δ b*, the difference between the lesion and control area, was useful to judge whether a bruise was recent or older. When it showed a minus value, the bruise was diagnosed as recent within 2 days. The temporal course of Δ b* changes may be a good indication of the time of bruising as measured in days, however, in actual forensic analysis situations, it is difficult to measure changes in the skin area on a living person over several days. It was hard to perform further age analyses on older bruises especially where only one or two measurements had been taken. On spectrophotometric analysis, clear differences of spectral reflectance curves in some wavelength regions were observed in recent, older and nearly healed lesions as compared with controls. The characteristics of new bruises had a peak point around 500–520 nm of the green area and absorbed light remarkably in the green, yellow and orange wavelength regions (approximately 520–640 nm). On older lesions, reflectance in the blue and one portion of the green wavelength region (440–520 nm) remained unchanged or further reduced before approximately ending similarly to the control curve. Nearly healed injuries showed slight reductions of reflectance over wide wavelength regions, but the curves were close to that of control skin. These reflectance graph patterns could cope with the temporal sequence of color changes of a bruise. As individual reflectance graph patterns depended on particular factors, such as the degree of bruising, it was difficult to correspond the patterns of all bruises to the date of bruising. However, if two measurements were taken at 1 or 2 days, their validity for classifying into recent, older and nearly healed bruises, might be shown by comparing patterns of both curves patterns including the control area. Some indices using reflectance values on certain wavelengths were not very useful to pinpoint the time of bruising. On spectrophotometric analysis, the evaluation of the age of a bruise needs to be done by a comparison of reflectance graph patterns in wide wavelengths in visible light regions.