There is a wide variety of approaches to whole farm models (WFMs) including purely biophysical models as well as models that combine biological and financial elements. Our study was motivated by the notion that researchers may benefit from guidelines on the choices they must make about modelling approach, when they are interested in moving to the whole-farm scale in order to understand the impacts of management changes on farm businesses. This paper reviews 53 studies published in Agricultural Systems, including the recent development of WFMs for developing country situations. We document current approaches and develop a typology, and describe strengths and weaknesses of various approaches. Models differed in the extent to which they accounted for resource constraints, endogeneity of input levels, spatial heterogeneity, interactions between activities, inter- and intra-annual variability, and risk. Models varied in objective (profit, environment, household food security), audience and whether they used real or representative farms. We found many studies did not provide a new insight into a farming system that has general relevance outside the specifics of the case being examined, did not use sensitivity analysis or validation, and were unclear on the objective of the work and the target audience. In response to this we consider issues around the communication of this type of research, proposing guidelines for the publication of WFM papers that are well documented, clear and useful.