An algorithm has been built, tested and compared for the problem of coordinated inventory replenisment of items having nondeterministic demands with time varying means. The objective is to minimize the total inventory cost, which is constituted by inventory carrying cost, ordering cost (or set-up cost in the case of manufacturing) and stock out costs. The algorithm calculates an index which represents the ordering frequency of each item. The item with the highest index is chosen and an optimum ordering sequence is established for this item assuming probabilistic and time varying demand. The other items are ordered only when this item is ordered. Families consisting of 5 to 25 items were tried. The algorithm generates a suggested ordering sequence and ordering quantities for all the items. Simulated demand was used to test the algorithm and to compare to Silver's approach. The new algorithm results in an inventory cost that is five to twenty five percent less than Silver's approach.