With the explosion of on-line non-English documents, cross language information retrieval (CLIR) systems have become increasingly important in recent years. Moreover, roles of individual entities in an information retrieval system, specially a cross language one, can be implemented as distinct agents. Many methodologies for modeling the multi-agent system have been proposed. In this paper, we use Tropos methodology for modeling a Cross Language Information Retrieval System. The methodology leads in defining four different classes of agents: Personal-Agent, User-Agent, Translator, and Retrieval. Key Words: Cross Language Information Retrieval, Multi-Agent Systems, Tropos, Methodology