Intrusion Detection Systems (IDS) is defined as a component that analyses system and user operations in computer and network systems in search of activities considered undesirable from security perspectives. Applying mobile agent (MA) to intrusion detection design is a recent development and it is aimed at effective intrusion detection in distributed environment. From the literature, it is clear that most MA-based IDS that are available are not quite effective because their time to detection is high and detect limited intrusions.This paper proposes a way of classifying a typical IDS and then strategically reviews the existing mobile agent-based IDSs focusing on each of the categories of the classification, for example architecture, mode of data collection, the techniques for analysis, and the security of these intelligent codes. Their strengths and problems are stated wherever applicable. Furthermore, suggested ways of improving on current MA-IDS designs are presented in order to achieve an efficient mobile agent-based IDS for future security of distributed network.