Many people interact with websites around the world on a daily basis. A massive amount of data is generated and this information could be very respected for the business in the area of ​​accepting customer behaviors. Web usage mining is a relatively independent, but not sequestered, category that primarily describes techniques that discover the user's usage pattern and attempt to predict their behaviors. Web usage mining is the field of data mining that deals with novelty and the study of usage patterns with the use of web log data. Specifically, web logs aim to advance web-based applications. User identification helps identify who accesses the website and which pages are viewed. If users have access to their information, it is easy to identify users. In fact, many users do not save their information. In fact, a large number of users access websites through an agent, many users use the same computer, the existence of a firewall, independent users use different browsers, etc. All the difficulties make this work very complicated and very difficult, to accurately identify each unique user. We may use cookies to track user behavior. But considering privacy, many users do not practice cookies, so it is necessary to find other methods to solve this problem. For users who use the similar computer or similar agent, how to find them? As presented in [9], it uses a heuristic method to solve the problem, if a requested page is not directly accessible by a hyperlink with some of the pages visited by the user, the experiential assumes that it There is another user with the same computer or with the same IP address. Doru Tanasa and Brigitte Trousse [4] present a method called navigation ...... middle of paper ...... consideration from a web server. It is intended to allow companies to use cookies to understand the behavior of visitors online. But, Check the convenience of the methods of controlling cookies present on the computer, they are frequently restricted by users. USER IDENTIFICATION BY REFERENCE LOG The method used here is as follows. The REFERER_URL parameter collected along with the access log and the site topology are used to design navigation trails for each user (Cooley et al. 1999). If after all the pages a new page appears which is not accessible from the previously viewed pages, a new user is anticipated. An additional condition by which a new user is expected is that, in a path of previously viewed pages, it appears that a page has already been browsed. This situation is very limited and imprecise. It does not receive repeated pages from the same user in the same session, which is very public in real life.