The fastness of a dyed fabric is currently estimated using empirical rules, and generally the same rule is used for a complete class of dyes, such as reactive dyes. Actually, by measuring the different colorimetric corrdinates of a fabric before and after the test, it is possible to calculate a fastness assessment, this measurement being more reliable than the classical visual fastness assessment. The relation between the dye concentration in the fabric and the shade of the fabric is the Kubelka-Munck relation, which determines the dyeing. By using the degradation kinetics of a dye during the fastness test, it is possible to calculate the concentration variation, and with the Kubelka-Munck relationship, the change in shade during the fastness test and then a fastness assessment. But this simple calculation can be used only if we assume that the degradation of every dye is independent of the other dyes. If this hypothesis cannot be assumed, the calculation is much more complex. We have studied the prediction of shade degradation during a fastnesstest for cellulose fabrics dyed with reactive dyes.