Abstract. This paper discusses on the cyber-physical system technology that assures the interoperability of manufacturing control system and equipment in intelligent manufacturing domain. This technology involves an information exchange and communication process, where broad manufacturing resources of production workshops in advanced manufacturing industries such as control system, equipment, personnel, and materials may freely enter or exit from the distributed self-organizing peer-to-peer network as independent message nodes. In this process, an improved filtering and routing algorithm based on topic and eigenvalue similarity is also employed to group and compress massive manufacturing messages in the light of semantics and message clusters for the purpose of intelligently identifying, positioning, tracking, monitoring and managing manufacturing resources. Based on the features of manufacturing process, this paper filters messages and implements end-to-end batch compression and transmission so as to ensure that the massive messages transmitted among different manufacturing resources in the intelligent manufacturing process are equipped with the advantages like real time, fault tolerance and high robustness when data validity is guaranteed.