This paper reviews the current state of thermal modelling for buildings, and describes developments in data modelling, object-oriented programming and specific research projects which indicate the likely course of future developments. The limitations of the data structures and processes of current programs are described, in the context of the forces which drive development. Emerging international data exchange standards, which are likely to be a major influence on future engineering software, are examined in the context of building modelling. The applicability of object-oriented programming to building simulation is explained. Finally, a number of research projects are described; some aim to facilitate the production of new simulation programs, while others aim to integrate a set of design tools sharing a common building description.