A kinetic model was proposed for describing the curd assembly of skimmed goats' milk during enzymic coagulation. The enzymic coagulation of milk was monitored using an optical sensor that measured diffuse reflectance (light backscatter) at 880 nm. The appearance of a shoulder, at low temperatures and protein concentrations, in the diffuse reflectance ratio profile after the inflection point of the curve (Tmax) appeared to separate the aggregation and curd firming steps. The diffuse reflectance ratio profile after Tmax was attributed to the overlapping of casein micelles aggregation and curd firming reactions. The developed kinetic model combined a second order reaction model to describe aggregation reactions and a first order reaction model to describe firming processes reactions. A completely randomised block design with three replications was used to determine the effect of protein concentration and temperature on kinetic constants. Milk was adjusted to three levels of protein (30, 50 and 70 g/kg), and coagulated at five temperatures (20, 25, 30, 35 and 40°C) to test a wide range of processing conditions. Data points from each profile after Tmax were fitted to the proposed model using non-linear regression. The average R2 and standard error of prediction (SEP) for 45 tests conducted were in the range of 0·9975±0·0027 and 0·0081±0·0037, respectively. A significant increase in characteristic times for aggregation (τ2) and curd firming processes (τ1) were found when temperature decreased or protein increased. Theoretical asymptotic value of reflectance ratio, R∞, increased with increasing level of protein and temperature (P<0·05). The parameter β1, which represented the fraction of diffuse reflectance ratio attributed to aggregation, decreased with increasing temperature and decreasing protein.