The objective of this study was to test the reliability of the post-mortem evaluation of bruise characteristics using a newly extended protocol. Intra-observer reliability of the observations was assessed with three meat graders, each of whom blindly examined 50 photos of bovine carcasses in two sessions, with 1 week in between. The agreement was “substantial” for number of bruises per carcass, and size and shape of the bruises; whereas it was “moderate” for colour and severity grade. Inter-observer reliability of the observations was assessed with four meat graders who each examined the same 46 carcasses at the slaughter-line. Inter-observer agreement for number of bruises per carcass ranged between observer pairs from “fair” to “moderate.” The agreement for number of bruises scored per anatomical site was “very low” for all observers. The most concordant bruise feature was size, followed by colour and shape. The bruise scoring protocol was associated with high level of intra-observer reliability and low inter-observer reliability. A further calibration of the observer judgments is needed to reach a more uniform scoring of the observers.