This paper considers a dynamic remanufacturing planning problem, in which wastes are collected by multiple container types and can be remanufactured at each period by any value in the set {0. P, …, mP} where the rate is the increment of a remanufacturing capacity and is a nonnegative integer. Each container type has a type-dependent carrying capacity and the freight cost is proportional to the number of containers types used. The multiple products are remanufactured by each taking a fixed portion (0 〈 αi 〈 1) of the input wastes to satisfy dynamic demands of each product over a discrete and finite time horizon. Also, a start-up cost is only incurred at the first period of a remanufacturing block which is consecutively remanufactured. It is assumed that the related cost (collection and inventory holding costs of the wastes, and the remanufacturing and inventory holding costs of the remanufactured products) functions are concave and backlogging is not allowed. The objective of this paper is to simultaneously determine the optimal waste collection and remanufacturing plans that minimize the total cost to satisfy dynamic demands of the multiple products. In this paper, the optimal solution properties are characterized and then, based on these properties, a dynamic programming algorithm is presented to find the optimal plan. Also, an acyclic network model is proposed to efficiently find the optimal solution to ()-subproblems. Finally a numerical example is introduced to illustrate the procedure for applying the proposed algorithm.