Multiagent systems have become increasingly important in developing complex software systems. Multiagent systems introduce collective intelligence and provide benefits such as flexibility, scalability, decentralization, and increased reliability. A software agent is a high-level software abstraction that is capable of performing given tasks in an environment without human intervention. Although multiagent systems provide a convenient and powerful way to organize complex software systems, developing such system is very complicated. To help manage this complexity this research develops a methodology and technique for analyzing, monitoring and troubleshooting multiagent systems execution. This is accomplished by visualizing a multiagent system at multiple levels of abstraction to capture the relationships and dependencies among the agents.