Log analysis or Web transactions Log are beneficial for enhancing information retrieval systems, productivity of the system for users, and understanding how users benefited from a system and their information behavior. The purpose of this paper is to analyze users’ behavior, the level of their using IRANDOC database and the way they use it. In order to do this, the users log to the database from outside IRANDOC searching information was studied for a period of 24 days in Bahman and Esfand 1391. The analysis of users log includes: geographic coding analysis, statistical analysis, and microanalysis of information seeking behavior. The results indicate that users from the cities of Tehran, Mashhad and Tabriz used the database more than other cities. Sundays and Mondays between 9 to13 were the peak use days and hours. 10 to 15 search characters long were the most frequently used search items, and the most searched subject was Social Sciences. The findings showed also that some users were not familiar with concepts like search term modification and how to use Boolean operators which resulted in inefficient information retrieval. Inefficient retrieval algorithm for short words was also found in the database. It was also found that log analysis can be of use in the evaluation of an information retrieval system as well as in user behavior analysis.