A kinetic modelling approach was developed and investigated with the aim of predicting the utilization of major substrates in the mammary gland and milk secretion rates in the lactating cow at varying concentrations of substrate in arterial blood. The model includes kinetic equations of transport and metabolism of glucose, acetate, free amino acids and free fatty acids in secretory cells and a phenomenological description of autoregulation of local blood flow, in which an energy criterion of control has been used. The predicted relationships between the rate of milk secretion and glucose levels in the blood are consistent with experimental results. Differential stimulation of α-lactalbumin synthesis causes increments in local blood flow and milk secretion rate in the model. The results of the study suggest that there is no simple relationship between the level of substrates in the blood and milk yield and contents of fat and protein in milk. This is because the effect on production of varying patterns of substrate concentrations in the blood is mediated by network interactions at the level of secretory cell metabolism and microcirculation. However, dynamic modelling provides a rational framework for developing such predictive tools.