In global supply chains, coordination of decision making in decentralized supply chain has received much attention. The total profit in the decentralized situation is inferior to the optimal profit in the centralized situation. In order to improve the total profit in the decentralized situation, it is required to optimize the coordination of the production planning for supplier and retailer under different leader-follower relation under demand uncertainty. In this paper, the multi-period production planning problem for the supplier and the retailer is formulated as the nonlinear bilevel programming problem under uncertain demand. We propose the replacement of leader-follower relation in multi-period supply chain planning. In the proposed model, the leader can be changed dynamically in the middle of periods. The effectiveness of the replacement of leader-follower relation to the total profit in the bilevel production planning is examined by some numerical experiments.