Abstract: A method for recommending bookmark data includes receiving a bookmark data set from a user terminal of a user, the bookmark data set including a main topic and a plurality of sub-topics, each of which includes a sub-title and a plurality of URLs; assigning a bookmark eigenvalue for the bookmark data set; and determining and transmitting another bookmark data set similar to the bookmark data set or another sub-topic similar to a sub-topic included in the bookmark data set to the user. The recommendation is made based on the bookmark eigenvalue, which is determined by assigning a URL eigenvalue for each URL included in each sub-topic by analyzing a web page that corresponds to the URL; assigning a sub-title eigenvalue for the sub-title of each sub-topic; and assigning a sub-topic eigenvalue in consideration of the URL eigenvalue and the sub-title eigenvalue.