Land diversity is a characteristic of low-input farming systems. Land diversity can refer to between-field diversity of grassland vegetation types that are a result of management practices (fertilisation, grazing, cutting), and environmental factors (altitude, aspect, soil water capacity) that have an influence on herbage production. Land diversity can also concern other characteristics of the field, like its distance to the cowshed or its suitability for mechanization for hay-making that are key characteristics affecting the spatial organization of management practices. The purpose of this study was to evaluate, using a simulation model, the consequences for yield of taking the land diversity at farm level into account in the decision-making process. We made the assumption that it could be an asset rather than a constraint for the management of the forage system, and that it could lead to an improvement in the forage balance in dairy farming systems. The model is able to simulate the decision-making process in managing this diversity during a growing season. The decision process was translated into three management strategies of the forage system representing increasing consideration of the farmland diversity. We then simulated these three strategies for two weather series and two farmland diversity levels. We noticed that the strategy giving the most consideration to the farmland diversity improves the forage self-sufficiency (due to higher yields and quality of forage supplies, and a decrease in the amount of hay fed during the grazing period). We thus concluded that the management of between-field diversity could be an asset for extensive farming systems. In the context of extensification of agriculture and increasing concern about its environmental impact on biodiversity, this approach must be pursued. The oriented-object structure of the model will facilitate improvements. © 2006 Elsevier B.V. All rights reserved.