Most of Disaster Managing Approaches depends firmly on information and communication management system. Due to the advances in Information and Communication Technology (ICT), a natural disaster, when occurs in an urban place, requires a high reliable quality of online heterogeneous urban information. In this context, decisions require the involvement of many decision makers and must be made in very short laps of time by using most of the available information. Since the information space has been opened the task of getting "online" the "more relevant information" has become real dilemma due to the huge structured and non structured available information whether on the Internet or on the Intranet. In our project a Disaster Decision Support System is based on a groupware approach in which a collaborative information board has to share within the different decision makers. Such system incorporates, at least and in addition to the interactive decision board, an optimised Data Base Management System, a light Geographical Information System, an Electronic Document Management System and a very powerful server. We consider that the information updating (as it is displayed or requested) on and within this decision board can enlighten a lot about the context. Consequently, we have been interested in Context Modelling available approaches. Most of these modelling approaches don't, exceptionally, fit to crises or disaster context. Therefore we built our own model based mainly on the approach of context ontology We have also developed a context capturing approach capable to retrieve the "more appropriate" and "efficient" "contextual" information (or web-document) to urban disaster management system. Every time the context is captured it will be used to generate continuously an orderly sorted set of requests.