This paper presents a DCOM-based framework for design of distributed scheduling systems and its application to flexible flow shop scheduling systems. We model two kinds of agents for each shop: scheduling agent and communication agent. The communication agent communicates with related agents dynamically on DCOM. It reasons environment encountered and recommends adaptive scheduling algorithm to the scheduling agent, and also announces the inside situation and requires the cooperation from related agents. An application of flow shop is presented to the validity of the proposed method that exhibits attractive performances in terms of autonomy and collaboration.