This paper explores a series of methods and models of urban land demand forecast, and these methods and models were
dived into the urban land quantitative forecast and urban land pattern forecast two types. In this instance, we designed a
model base system framework, including the following steps: data processing, forecasting, forecast type select, forecast
way select, model checking and multi-schema evaluation, results output. On the basis of the basic flow, object-oriented
analysis technique was adopted for the developments of prediction models and prediction flow, and then a model base
system with discreteness characteristic were produced, which was extendable and reusable. Finally, based on Web
Service technique, we designed urban land demand model base system framework, which provides an integration
scheme for the GIS service and application services.