Abstract:With the advent of the era of big data, the information generated in various fields is growing rapidly. Some of them have the characteristics of large amount of data, real-time arrival, continuous uninterrupted, fast change, etc., and need to be processed in real time. The processing requirements of these data are also complex and changeable. The business personnel need to carry out corresponding algorithm customization, which not only requires relevant programming knowledge, but also has to deal with cumbersome processing flow and lengthy development cycle. In order to solve the above problems, this paper designs and implements a stream data processing and service system based on process modeling, which provides real-time access to multi-source stream data, stream data service and stream data processing service. The system encapsulates the stream data processing process into a service provided to the user, allows the user to drag and drop the combined stream data processing and service module, configures relevant parameters, defines the process of stream data processing and service, and implements stream data processing quickly and naturally. And service-oriented tasks, the processing results are pushed to other application systems in real time through service routes to meet different business needs. Case studies show that the system is more efficient, flexible, and configurable than traditional streaming data processing systems, and has advantages in terms of usability, usability, and scalability.