The distributed smart factory platform provides an environment where various devices deployed in the factory can perform intelligent collaboration while horizontally interacting with the devices around them. This requires individual devices to be able to process and manage their own data stream collected in the shop floor. In this paper, we focus on building a system for efficiently managing data generated in a smart factory environment, and propose a data stream management structure to support it. In the proposed approach, sliding window and continuous query processing mechanisms are introduced to immediately process various types of data stream collected from devices at the time of occurrence. In addition, we propose a strategy to store processed historical data in local database by introducing temporal approach.