Prepacks are often provided by suppliers for the promotion in retail distribution, which groups multiple units of one product or combinations of related products together with lower price. The prepack option can reduce retailers' procurement cost, whereas may limit the ordering flexibility and increase supply-demand imbalance. This paper studies the dynamic inventory planning problem for an online retailer that implements a two-layer distribution network consisting of regional and forward distribution centers (FDCs). At the start of each period, the retailer decides how to replenish inventory using both prepack and individual replenishment options, and how to allocate the inventory of each product to FDCs. At the end of each period, the retailer chooses from which FDCs to fulfill the demand. We formulate a multi-period mixed-integer stochastic optimization model and propose a two-phase approach to solve it. Phase one determines the integer prepack replenishment decisions using a target-oriented robust optimization approach, phase two applies a linear decision rule to adaptively determine the individual replenishment, allocation, and fulfillment decisions. Numerical results suggest that our approach produces good-quality solutions efficiently and using the prepack option creates significant values. A case study using real data from JD.com demonstrates the applicability of our approach.