This paper researches the optimization of inventory system with discrete time, single product and single period under uncertainty environment. The aim is to analyze the impact of the alterability of the initial storage, production batch, output and need on system optimal policy and optimum cost. According to random order, we put forwards the random comparison result of optimal inventory and optimum cost under single period environment. Finally, a case study is given to show the feasibility and operability of the method, which can provide basis for supply chain managers making effective decisions.