Scheduling of production plants is often affected by random side conditions such as random demand or process-inherent uncertainties. We address multiproduct batch plants where small lots of specialized products are manufactured for fast changing markets. A two-stage stochastic programming model for the production planning in a multiproduct batch plant is introduced. The model is transformed into a large-scale deterministic mixed-integer program, for which we propose a decomposition method based on Lagrangian relaxation.