The new trend in network computing involves using mobile agents to implement distributed applications. A mobile agent is an autonomous process acting on behalf of a user and roaming the internet to access data and services. In order to show the performance as well as execution status of a mobile agent system, an on-line system monitor is required. Monitoring technique gathers information about a computational process as it executes. An on-line monitoring system is an external observer, monitors a fully functioning application or system, and is generally intended to be permanent. This thesis discusses the design strategy and the implementation of the On-line Monitor in IMAGO (Intelligent Mobile Agent Gliding-Online) system, which is an infrastructure for mobile agents. The IMAGO On-line Monitor captures run-time data from a set of mobile agent servers, and analyzes those data to gain insight into the status of servers as well as their hosted agents. Two models for data collection will be discussed and compared, where one model uses probing strategy and another adopts push strategy. The implementation of the On-line Monitor provides a user-friendly graphic interface to show running status of mobile agents and related statistics of the whole system.