Traditional aquaculture fish ponds are located outside people’s living areas. Collecting water environment data and uploading them to the cloud server for processing requires high network resources. Using edge computing power to process local data can solve network resource problems while achieving intelligent coordination between devices. Using the Raspberry Pi as an edge computing device in the local space, the system is explained from the three main aspects of the perception layer, the transmission layer and the application layer. The goal of real-time and refined resource management is achieved through intelligent algorithms, so that the data is respectable and credible. Further explore the feasibility of intelligent collaboration in aquaculture. Practice shows that the system runs stably on the physical model and has the characteristics of fast response in perception and control.