nutritional requirements of beef cattle described by nrc (2000) are determinated by equations, some of which are nonlinear or non additive which makes it difficult for modelling with linear programming. this paper describes a model of mixed integer linear programming to formulate least-cost diet that incorporates the recommendations of model level one of nrc (2000). the lack of additivity or linearity of some functions were overcome using alternatives restrictions modelling with binary variables. the model considers environmental factors that modify the requirements of the animals and the effects of the diet that affects the efficiency of microbial protein synthesis in the rumen and the level of voluntary feed intake. this model considers the requirements of energy, metabolizable protein, calcium, phosphorus and effective fiber. the intake of pasture is limited by the potential intake of pasture at grazing and the intake of supplement is limited by the maximum level of supplementation permitted. this paper shows an example of application of the model and describes possible extensions.