Agenda 2000, the reform of the CAP (Common Agricultural Policy) in progress, implies significant mutations in the technical and economical environment of French farms. Therefore, an important question for the civil authorities is: 'What would be the impact of such changes on cultivated surfaces?' To this end, a micro-economic model, whose purpose was to study the regional agricultural supply, was built by Sourie et al. at INRA-ESR (National Institute of Agricultural Research, Rural Economy and Sociology, Grignon, France). This model, named MAORIE - the acronym in French for Regional Agricultural Supply Model INRA Economy - is a Linear Programming (LP) model intended to represent the farmers' behaviour as to their surface allocations to various cultures. The aim of the present study is to improve the representativity of MAORIE by taking into account uncertainty about crop prices and yields. To model this uncertainty, intervals on gross margins per surface unit were introduced in the objective function level of the model. The resulting model is an Interval Linear Programming (ILP) model. Many resolution procedures have been proposed for ILP models in the literature. In this work, the minmax regret approach will be investigated to determine whether it can lead to the desired representativity.