Abstract: Under the urgent requirements of information development,research on data warehouse is now raising and becomes a new international frontier research direction.Firstly,the paper analyzes some problems in traditional urban planning decision support system,that is:(1)This system can only provide with data support of aided planning decision,but without semi-structured decision problems if traditional database techniques are used.(2)Integrated data are the basis of decision support system,however,it is not easy to integrate data which distribute in varied heterogeneous data platform.(3)If the decision support system is established on the basis of database management system,the ability of spatial problem of knowledge expression and knowledge synthesis will be not very strong because of weak calculation ability of database.Secondly,the paper presents preliminarily system architecture of urban planning decision support system based on data warehouse.The system consists of database,planning data warehouse,database warehouse management module,data mining tool,knowledge base,knowledge discovery module and user interface.The key problems are how to create planning data warehouse,how to discover knowledge from data warehouse,and how to apply knowledge discovered to the practice.Thirdly,the paper discusses organization of planning data warehouse,data mining techniques and methods of knowledge discovery.Data in urban planning and management come from basic geographical information,planning results,and so on.The planning data warehouse can be divided into three levels: historical data level,synthesized data level and special data level.The historical data level is used as urban trend analysis and urban simulation.The synthesized data level includes varied statistical data,planning control index,evaluation results and forecast analytical data.The special data level stores data of population,resource,economics,environment,and some parameters in mathematical models,prediction models and evaluation models.The planning data warehouse can be built with multi-dimensional database.The key problem of establishment of whole data warehouse is the organization of spatial dimensional data.Data mining is just knowledge discovery on data warehouse.It can apply to decision support.It is advantageous for knowledge discovery gradually when the level spaces of data mining are established.The process of knowledge discovery is the process of decision support.Comparing with information from database management system,the knowledge from data warehouse should be hidden,refined and high standard.In data mining,knowledge can be regarded as concepts,rules,regulars,modules,restrains and visualization.After explanation,these knowledge can be applied into practical system.The paper also presents some methods of knowledge discovery.On the above discusses,a method of establishment of urban planning decision support system is proposed in the paper.Finally,taking an example of Jingzhou Environmental Information System,it presents system application under the support of urban planning decision support system.