The evolution of the inventory level under stochastic supply disruptions and returns no longer varies mono-tonically but fluctuates stochastically, which makes it very difficult to control the inventory level. In order to solve the inventory shortage and overstock problems, a contingent control (including contingent sourcing and contingent disposal) policy is proposed in this paper. Under the condition that the inventory level process is expressed as a L′evy process, the expected total discounted profit model is derived by utilizing continuous-time Markov chain, renewal process and martingale theorems. Subsequently, the cross-entropy method is designed to obtain the optimal control policy. Numerical results show that the intensity and the types of disruptions, as well as the arrival rates and the batch sizes of returns are critical determinants of the optimal contingent disposal level and contingent sourcing size. However, the types of returns have big impacts on the optimal