Effective inventory management across a supply chain is very important for reducing inventory costs while improving services to customers. One problem for the management is to determine all optimal inventory policy for each stock in the supply chain. The problem is difficult to solve not only because a supply chain is a multi-echelon inventory system with multiple interrelated stocks but also because it involves conflicting objectives. Finding a set of pareto-optimal solutions for the problem requires a robust and efficient method Chat can efficiently search the entire solution space of the problem. Genetic algorithms (GAs) seem to be suited for this task because they process multiple solutions in parallel, possibly exploiting the similarities of the solutions by recombining them. In this chapter, supply chain inventory polices are optimised using a multi-objective optimisation approach that combines a genetic algorithm with a Petri net-based simulation toot for performance evaluation. The supply chain considered is first modeled as a batch deterministic and stochastic Petri net, and a simulation-based optimisation method is developed for parameter optimisation of inventory policies of the supply chain with a multi-objective optimisation approach as its search engine. In this method, the performance of a supply chain is evaluated by simulating its Petri net model, and a Non dominated Sorting Genetic Algorithm (NSGA2) is used to guide the optimisation search process toward high-quality solutions. An application to a real-life supply chain demonstrates that our approach can obtain inventory policies better than ones currently used in practice in tern-is of inventory cost and service level.