Embodiments of the present invention provide a method for processing a World Wide Web (WWW) page, which includes: determining at least one website to be optimized; generating a corresponding page template for each of WWW pages with different types in each website, and storing the page template; grabbing WWW pages from each website, matching each grabbed WWW page with a page template, filtering redundant HTML data from the WWW page according to a matching result, and storing the filtered WWW page; after receiving a request sent by a terminal for accessing a WWW page, determining whether there is a stored filtered WWW page corresponding to the WWW page requested by the terminal, if yes, returning the filtered WWW page to the terminal. Embodiments of the present invention also provide an apparatus for processing a WWW page. With the scheme of the present invention, redundant information may be efficiently eliminated.