The objective of our study was to explore relationships between Japanese color score (JCS) and pork-quality attributes and develop equations to predict JCS. Pork carcass traits in population one (n = 781) was used to develop prediction equations and population two (n = 684) was used to test the equations for accuracy. Pearson's correlation coefficients found firmness, ultimate pH, drip loss percentage, L*, a*, b*, hue angle, and chroma were significantly (P < 0.01) correlated to JCS. Correlation loading found 96% of the variation in firmness, pH, drip loss percentage, L*, a*, b*, and hue angle explained 81% of the variation in JCS. Three prediction equations were developed using these traits. Averages for population one traits were used to develop the initial prediction equations. Predicted JCS, which fell within [plus or minus]0.25 of the actual JCS, were retained and multiple linear regression (MLR) was run, resulting in the first prediction equations. Data from population two were then used to evaluate the success of these equations. Equation one using firmness, pH, drip loss percentage, L*, a*, b*, and hue angle was: JCS = 12.698 – (0.00007199 x drip loss) + (0.09008 x pH) – (0.01128 x firmness) – (0.226 x L*) + (0.06923 x a*) – (0.0201 x b*) + (0.02143 x hue angle); r[superscript]2 = 0.916. For the test population, 98.53 and 67.25% of the observations were predicted within [plus or minus]0.50 and 0.25 of the actual JCS, respectively. The second prediction equation, developed utilizing instrumental color measures of L*, b*, and hue angle was: JCS = 15.255 – (0.259 x L*) – (0.213 x b*) + (0.02518 x hue angle); r[superscript]2 = 0.931. For test population, 92.40 and 55.70% of the observations were predicted within [plus or minus]0.50 and 0.25 of the actual JCS, respectively. The third prediction equation developed utilizing L*, a*, and b* was: JCS = 12.920 – (0.219 x L*) + (0.07342 x a*) – (0.02166 x b*); r[superscript]2 = 0.906. For test population, 97.80 and 68.22% of the observations were predicted within [plus or minus]0.50 and 0.25 of the actual JCS, respectively. All prediction equations predicted 92% or more of the JCS observations within [plus or minus]0.50 and would be useful when sorting pork carcasses for export to valuable Asian markets. The second and third prediction equations would be advantageous as they require fewer measurements and could be more rapidly collected.