Optimisation models are a key tool for the analysis of emerging policies, price sets, and technologies within grazing systems. A detailed nonlinear optimisation model of a New Zealand dairy farming system is described. The framework is notable for its rich portrayal of pasture and cow biology that add substantial descriptive power to standard approaches. Key processes incorporated in the model include: (1) pasture growth and digestibility that differ with residual pasture mass and rotation length, (2) pasture utilisation that varies by stocking rate, and (3) different levels of intake regulation. Model output is shown to closely match data from a more detailed simulation model (deviations between 0 and 5 per cent) and survey data (deviations between 1 and 11 per cent), providing confidence in its predictive capacity. Use of the model is demonstrated in an empirical application investigating the relative profitability of production systems involving different amounts of imported feed under price variation. The case study indicates superior profitability associated with the use of a moderate level of imported supplement, with Operating Profit ($NZ ha-1) of 934, 926, 1186, 1314, and 1093 when imported feed makes up 0, 5, 10, 20 and 30 per cent of the diet, respectively. Stocking rate and milk production per cow increase by 35 and 29 per cent, respectively, as the proportion of imported feed increases from 0 to 30 per cent of the diet. Pasture utilisation increases with stocking rate. Accordingly, pasture eaten and nitrogen fertiliser application increase by 20 and 213 per cent, respectively, as the proportion of imported feed increases from 0 to 30 per cent of the diet.