In forest fire protection systems 24 hour surveillance is of most importance. Real time data must be acquired to react fast enough to suppress fire occurrence or to minimize damage made by forest fires. In forest fire monitoring systems, usually a large area must be controlled. Real time data has to be collected and processed in time. When the amount of data reaches critical volume, modern software techniques have to be implemented in order to accomplish system goals. In this work we have implemented agent technology on data retrieval and processing. A multi-agent system for real-time data collection and processing is described. This work is a part of a more complex integral project of forest fire protection in Split-Dalmatia County. The integral forest fire protection system will be based on the information system for integration of all activities connected with early fire detection by 24 hours video and meteorological monitoring, management of forest-fire fighting and post-fire recuperation of burned landscape

By appropriate integration of advanced ICT systems in an environment, in a way that those highly technological systems become almost its integral part, it is possible to provide additional environment features. Those features are primarily self-monitoring and self-protection, giving the environment rudimentary intelligence and possibility, to operate not only by reaction, but also to operate proactively, having in "mind" its self-protection. In such a way the environment becomes the intelligent environment or more accurately the intelligent self-monitoring, self-protecting and self-aware environment that reacts on changes and in real time alarms humans responsible for appropriate environment protection actions which will prevent environment further degradation. The paper describes overall architecture of such intelligent environment based on advanced sensor network called the observer network. As an example the system architecture of the forest fire monitoring system is discussed.