This article deals with a single perishable item, continuous-review, two-echelon serial inventory system consisting of a warehouse and a retailer. Customer demands at the retailer are assumed to be Poisson. All items have a xed shelf life and start aging on their arrival at the retailer. The demand that cannot be met immediately at the retailer is lost. All transportation times are xed. If there is any stock in the warehouse, the lead time for the retailer would be the transportation time from the warehouse. Otherwise, the retailer orders are met with a delay. In this article, using an approximate technique, we rst present a heuristic for nding cost-eective base stock policy and then develop a simulation-based neighborhood search procedure to modify the quality of the solution. Also, a numerical experiment is carried out to evaluate eectiveness and accuracy of the procedures. The results reveal that the approximate model performs rather well.