Abstract: Extensive simulations and good scheduling rules are usually necessary to obtain the optimal schedule in the simulation-based scheduling methods. We present a methodology to build a decomposable model for scheduling the semiconductor manufacturing systems, with the objective of minimizing the average WIP. The model is decomposed and further simplified in the context of semiconductor manufacturing. The conclusion can be directly applied to simulation-based scheduling as a scheduling rule. Because it uses the global state information in decision, the optimization ability and performance of simulation based scheduling are improved without increasing the simulation runs.