Decision making studies in agriculture are often made difficult by the complex and dynamic nature of bio-economic systems. Simulation is one of the newer systems research techniques which as yet has had limited use in farm management research. This paper discusses some methodological aspects of simulation with specific reference to grazing systems. Problems arising in the development and use of simulation models are discussed and the need for inter-disciplinary co-operation to overcome data problems is indicated. One approach to experimentation is illustrated by reference to a model of a sheep grazing system and the problem of cropping for winter grazing. It is concluded that simulation is a potentially useful technique for management-oriented systems research in agriculture.