A service-oriented approach for an open, scalable and generic River Information System (RIS) is presented. In a brief introduction the requirements for such a distributed environmental information system (EIS) in this special field are given. They define the basics for an adequate conceptual and technical system design, based on WWW technology and use of CORBA services. It is shown how and in which way public access and internal administrative data distribution restricts availability of data and system structure. The developed strategic approach leads to an overall system, which is transferable to other EIS. Embedding future needs to provide an elegant solution to incorporate ODBMS access and entries to workflow management systems (WFMS). Division in several layers guaranties a maximum of flexibility and scalability. Use of CORBA services drives the evolution of this EIS to a decentralised data collection, storing and updating, which is essential for long-term efficiency and user acceptance (keep data where it is collected). Data selection, handling and operations on data are performed by intelligent agents, operating on specified methods in an authorised environment (e.g. server, data, methods and user). The technological architecture of the system and the implementation of selected components and features are described. It is pointed out why such a system must be adaptable to present and future commandments and how this can be realised under controlled conditions for a self-growing system. The complex and sensitive handling of data is highlighted by an example. Some possible lines for WMFS access are given. The potential for automation, quality assurance and controlling is shown. To date, a comfortable solution has been developed, that can be used by decision-makers, administration, industry, scientists and the general public to have access to all environmental data related to rivers. The developed system can be adopted to EIS for estuaries, channels for inland navigation and coastal regions.