In make-to-order systems, customers expect more freedom to choose the accessories they desire. However, demand variations and internal disorders cause uncertainties. Hence, a different inventory system is required for such items to dynamically manage those variations. In this paper, a dynamic approach based on the theory of constraints is proposed for inventory planning and control of accessories. First, the risk of processing time variation is balanced while keeping cycle time balanced. Second, the ribbons of buffer control charts are determined by a buffer planning model in which a multicriteria ABC analysis is carried out to apply different customer service levels. To detect demand variations and monitor the buffer, trend of consumption in each monitoring window is carefully traced. In addition, simulation-based procedures are recommended to update control ribbons. The comparison between the performance of the proposed approach and those of common methods using the data of an automobile company, as well as several random test problems, confirms that the total cost and the efficiency of inventory system can be significantly reduced and increased, respectively.