Abstract : We present the on-going research project ISIS (Interoperable Spatial Information Systems) which is a semantic mediation approach to support geographical spatial information systems interoperability. The diversity of spatial information systems (GIS) and data models has created a need for tools and methodologies to allow cooperation or interoperation among GIS. In particular, mediation based approaches, which aim at achieving autonomy, extensibility and flexibility of participant systems, will become increasingly popular in emerging WEB-based processing environments. The underlying approach of ISIS combines techniques from the domains of traditional interoperable information systems, spatial data modeling and multi-agents systems. In this paper we describe the Amun data model which provides a foundation for resolving semantic discrepancies among systems and for modeling cooperative GIS.