The research aims to model wine producer behavior and to assess the effectiveness of policies to reduce pesticide use in winegrape growing. We model the decision of a producer maximizing expected income under constraints and phytosanitary risk, considering the impacts of different strategies targeting downy mildew, the main grape disease in European Atlantic vineyards. The VINEPA model is a multi-periodic stochastic programming model based on panel-data of about one hundred representative winegrowing farms from the Farm Accountancy Data Network in the Bordeaux region. The response of vine to fungicide treatments against downy mildew is simulated through Downy Mildew Potential System, an epidemiologic model initially developed for decision support, using data from multiple weather stations and untreated vine plots weekly monitored over a ten years period. The model accurately reproduces the current chemical protection strategies in the region. Simulations are then carried out for different types of taxes (ad valorem and volume based), with increasing rates. Effects of policies on spraying practices are analyzed along with their potential impact on investment in precision technology equipment. Shortcomings and further developments with applications of the VINEPA model are discussed.