Grazing systems constitute the most extensive land use worldwide. However, economic analysis of these systems has mainly involved the use of linear optimisation methods that provide a general description of the complex processes contained therein. This paper describes a nonlinear optimisation model of a New Zealand dairy farm that incorporates a detailed depiction of key biophysical processes present within grazing systems. The capacity of this optimisation model to provide rich insight into the effects of higher stocking rates within grazing systems is demonstrated in an empirical application. In accordance with system trials, this application shows that higher stocking rates on pasture-based New Zealand dairy farms generally increase pre-grazing pasture biomass, decrease post-grazing pasture biomass, increase pasture utilisation, decrease herbage allowance, decrease intake and energy consumption per cow, decrease milk production per cow, increase milk production per ha, and reduce conception rate. Nevertheless, an intermediate stocking rate is optimal, as greater milk production with a higher stocking rate is not sufficient to offset the associated costs. © 2013 Elsevier Ltd.