The knowledge on the Web is increasing rapidly. Without using a recommendation systems, the users have spent maximum time on the Web to get the data. So, a recommendation system is very important. But, now a day’s many recommendation systems cannot give the exact information to the users. The reason is they cannot extract user’s interests accurately. So, analyzing the user’s interests and identifying the correct domain is important research in Web Usage Mining. If users' interests can be automatically detected from users' Web Log Data, they can be used for information recommendation which will be useful for both the users and the Web site developers. In this paper, one novel algorithm is proposed to extract users' interests. The algorithm is based on visit time and visit density. The experimental results of the proposed method succeed in finding user's interested domains. © Research India Publications.