Information monitoring is a very essential activity for management, control and decision making in various applications. This paper proposes an intelligent system for monitoring information sources that consists of many adaptive agents cooperate with each other in order to monitor every information item updates at the right time without over loading of system resources. The main key of the presented system is the using of data mining approach to find information updates depending on the update behavior of each item and the relations between items updates. An experiment to investigate the performance of the proposed system is conducted by simulating real-world operations of financial management in merchandise trading transactions. Experimental results show that the proposed system detects more updates in a timely manner by incurring less extra loading to the related information servers and the communication channels.