Omni-channel retailers meet the demand through the e-commerce channel and the stores in different locations. They have specific centers to fulfill their online orders (e-fulfillment centers); however, they may fulfill their online orders using their in-store inventory. On the other hand, retailers can use pricing decisions to control the demand in each channel. We first model the customers’ demand by the multinomial logit (MNL) choice model. We then propose two mixed-integer nonlinear programming (MINLP) models to maximize the total profit by considering both fulfillment and pricing decisions. We assume e-fulfillment inventory can also be used as a backup inventory for in-store orders if stores are out of inventory, but only for a store in the same zone. We also assume limited initial inventory and no replenishment in a finite selling horizon. In the second MINLP model, we also allow for multiple shipping options and inventory decisions. By numerical experiments, we show that omni-channel retailers can increase their profit by having control over initial inventory allocation and fulfilling the in-store demand by e-fulfillment inventory.