Environmental emergencies such as forest fires present a serious threat to the environment and human life. The monitoring system of emergencies should be capable of early detection in order to reduce possible damage. In this paper, we present a new approach for forest fire detection based on the integration of Data Mining techniques into sensor nodes. The idea is to use a clustered WSN where each node will individually decide on detecting fire using a classifier of Data Mining techniques. When a fire is detected, the node will send an alert through its cluster-head which will pass through gateways and other cluster-heads until it will reach the sink in order to inform the firefighters. The proposed approach is evaluated using the CupCarbon simulator. The simulation experiments show that our approach can provide fast detection of forest fires while consuming energy efficiently.