In order to reduce pesticide use in vineyards, we propose a bioeconomic model to evaluate different fungicide treatment strategies. This model estimates the development of the downy mildew Plasmopara viticola for a given year’s weather on a grapevine plot, and predicts the damage done, the yield loss, and the resulting partial gross margin, depending on the chosen protection strategy. Grapevine growth and phenology are simulated with the STICS grapevine crop model according to the year’s weather; fungal components quantify downy mildew development; damage onto leaves and fruits is characterized as a percent reduction of potential leaf area and yield; the effect of fungicide treatments is simulated as a partial protection against infection; the economic result is calculated at plot level, taking into account simulated yield, local economic conditions and costs of observations and sprayings. The model parameters were estimated using three sets of experimental data from vineyards in the French wine-growing region of Bordeaux. Using these parameter values, the model was used to evaluate the following five protection strategies: 3 systematic fungicide spraying strategies with 2-, 3- or 4-week intervals, the “Mildium” adaptive strategy, which includes field observations and decision rules, and a control untreated strategy. Yield losses and the resulting partial gross margins were calculated for 23 annual weather examples for each strategy and the statistics of these strategies were compared. The adaptive strategy was found to be slightly less protective on average against downy mildew than the 2- week systematic spray strategy. However its low variability ensures sustainability in terms of grower’s income, while reducing by one third the number of sprayings. The model hypotheses and simulation results are discussed, in relation to the particular economic context of the “Bordeaux” protected designation of origin with the objective of reducing pesticide use in vineyards.