Most attempts to predict the effects of climate change on soils, and hence land use, have been made at coarse scales and have made little use of the detailed soil, land use and climatic information available. Much of this information is available in digital form and lends itself readily to manipulation by computer procedures, often within geographic information systems. The ACCESS project was designed to take advantage of this situation. The target was a spatially distributed soil, agro-climatic and soil hydrological model to predict the effects of climate change on land use within the European Community. In the event, the project was extended successfully to Hungary, Poland and Romania, which gave a much wider range of soil, soil hydrological and climatic regimes than originally envisaged. The model structure drew on earlier work, which related simple soil properties, such as might be obtained during soil surveys, to crop suitability. More powerful approaches to the estimation of the soil hydrological state and crop water demands were incorporated into the new model, as are new approaches to land type classification. Much effort also went into deriving robust pedo-transfer functions, which allow the derivation of soil hydrological properties from simple soil survey data. The new model (ACCESS) was purposely designed to run at two levels; a more general approach to utilise the results of the site specific approach to allow extrapolation of the modelling to large areas of land, and a detailed approach to use site specific data for calibration and validation. The most significant difference between the two routes through the model, is that the site specific model operates at a daily meteorological time-step, whilst the broad scale model runs at a monthly time-step.