Competition and resultant complexity of today's production industry require that enterprises realise the importance of reducing total costs in production. Uncertainties of customer orders also make it difficult to determine the schedule for minimizing inventories, earliness and tardiness. In this paper, we propose a new production model for flexible flow shops, which aims to reduce the total inventories, earliness and tardiness. The model divides a flexible flow shop (FFS) into a make-to-stock (MTS) part and a make-to-order (MTO) part by applying a decoupling point. In the MTS part, jobs are manufactured into semi-finished products and then stored as inventories at the decoupling point. As soon as a customer order is received, the inventories are released into the system, starting undergoing processing in the MTO part. This shortens the lead time for manufacturing the products. Another advantage of the proposed model is that the less number of operations in the MTO part than that in a job makes it easier to avoid tardiness. In addition, we designed two types of models: dynamic and static, which depends on whether the decoupling point is dynamically adjusted to adapt to different arrival rates of customer orders. The reason why we design two types is to compare the performance of the proposed two models. Results show that the dynamic hybrid model outperforms pull, push and static hybrid models for reducing costs.