We propose an implementation method of parallel production systems on message passing machines to speed up their execution time. There have been proposed parallel production systems using hash mechanism, but they cause skewed load distribution problem. To obtain more efficient balance of load, our method, named {\it CPPS} (Clustered Parallel Production System), adopts two load balancing strategies: hash and demand-driven. We estimate the execution time of the CPPS and simple hash method, and implement them on an nCUBE/2. The estimation meets the execution results. the CPPS method provides much better load balance to improve scalability.