Designing intercrop management plans (IMP) to meet objectives related to both crop production and environmental impacts is a challenge for farmers. A multiple criteria decision analysis is thus needed to evaluate and rate various cropping systems in different soil and climate conditions. Different intercrop management plans in vineyards were analysed in this study so as to classify them in relation to their ability to fulfil a particular set of objectives and to deal with climatic variability.The method included five steps. A set of intercrop management plans was defined by combining the type of grass, the covered soil surface ratio, and the intercropping duration. Four evaluation criteria were chosen: grapevine vegetative development, yield, product quality and runoff. Corresponding indicators were identified and the range of values that would be desirable or not were defined. A water balance model designed for row crops was run to simulate the behaviour of the grapevine–intercrop–soil system under different management plans and at various soil depths. The model was used to calculate, for each management plan and soil depth, the four indicators and evaluate the overall agreement and discordance with various weights assigned to the four criteria. A frequency analysis on 30 years of weather data was carried out to estimate the robustness of the most satisfactory intercrop management plans.The most satisfactory intercrop management plans differed according to the priority given to managing production or reducing environmental impacts and depended on the soil depth. This confirms the conclusions drawn on the basis of various experiments assessing a limited range of intercropping policies. Overall, giving priority to the environmental criterion was favourable for cropping systems with a long intercrop period regardless of the soil type. Few management plans were suitable for all years. The observed yearly rainfall varied from 330 to 1200 mm during the 1975–2003 period, which generated marked variations in the water balance of the grapevine–intercrop–soil system.The lack of robustness of the explored intercrop management plans could be a consequence of the poor description of the management plans due to the limited combinations of technical options considered. Strategic or tactical adjustments could be introduced by farmers. We assume that more robust intercrop management plans could be developed by introducing such rules in a decision model combined with the present biophysical model.