The target of the thesis is to investigate the dynamic pricing policies in a two-echelon supply chain in the presence of dynamic factors such as cost learning, demand saturation and reference price effects. We present Stackelberg differential games in a disintegrate channel consisting of a manufacturer and her independent retailer. Employing the optimal control and game theories, we derive the equilibrium pricing strategies of the manufacturer and the retailer and investigate the impacts of such dynamic factors on the channel efficiency. (l) First, we focus on the cost learning effect in the supply side. In the open-loop case where the manufacturer commits about the wholesale price preliminarily, a short-term price strategy is suggested to be optimal. Although each member benefits from a higher cost learning effect, the channel inefficiency is deteriorated. We highlight that, to mitigate the channel inefficiency, the manufacturer could either employ multiple competing retailers or renegotiate with the retailer about the wholesale price frequently. (2) Then we focus on the reference price effect in the demand side, which describes how past price affects current demand of the frequently purchased product. Comparing with the integrated channel, we indicate that price penetration is more preferable than price skimming regarding to a certain group of customers. Another finding is, the channel efficiency is always lower when the consumers are more sensitive to the gap between the real market price and reference price or have strong loyalty to the product. (3) What's more, we consider the joint dynamic effects in both demand (demand saturation) and supply (cost learning) sides in the stochastic supply chain. In the presence of decreasing uncertainty, the result demonstrates that demand uncertainty pushes up the retail price. Furthermore, numerical study illustrates that a deterministic pricing strategy could be better off than a stochastic pricing strategy in the decentralized channel. This counterintuitive result could be interpreted as that, the deterministic policy leads to underpricing the product, which exactly counteracts the overpricing problem due to the well known double marginalization effect.