A simulated annealing algorithm for batch process planning is described. The fortmulation treats a broad range of problems involving a general profit function with discrete demands and due dates. It is targeted to the planning of a batch process operated in "open shop" mode with minimum inventory and maximum on-time performance. The approach is demonstrated in case studies of a parallel network flowshop or "multiplant" structure with a zero-wait storage policy between stages. The effects of setup time and setup costs on the schedule and production capacity are illustrated in several case studies. A large and general case analogous to a real industrial problem is included. The approach finds significantly better solutions than the first come first serve, shortest processing time, or earliest due date heuristics often used for production scheduling.