This paper presents three new inventory Control models for perishable items in three-level supply chains using just-in-time logistics. The goal is to minimize the total cost of the whole supply chain. This includes cost of production, cost of freight, cost of preservation, cost of early or late delivery, and cost of perished goods. The first model, for the purpose of simplification, has been developed with the assumption that all goods are delivered to the customers prior to their expiry date and hence there are no perished items. To develop a more realistic model, the second model includes cost of perished goods and considers cases where some of the goods are expired prior to delivery to customer. The third model, developed to improve the second model even further, considers cases where close-to-expiry items are sold at a cheaper price. Genetic Algorithm and CPLEX are used to solve and validate the models.