Abstract: A user preferred browsing paths algorithm is proposed，and also be meaningful improved.The algorithm after improve：Firstly，two URL-URL matrixes are set up from Web logs according to Web site’s browsing paths，where referrer URL as rows，navigating URL as columns.Respectively，using the average visit times（divided by the number of characters in the website） and the frequency of visits as matrix element.Then，preferred browsing sub-paths will be discovered from the computation of this matrix．Finally，all the sub-paths are combined to generate a set of user preferred browsing paths.