Abstract: In a website, various information is organized by its own navigation system, which involves the semantic characteristics of classification. In order to fulfill effective extraction of Web information, the WCSW (website classification system wrapper) based on HTML page blocking algorithm is proposed aiming at the classification system of websites. WCSW deals with navigation information blocks involving semantic classification in accordance to extraction rules, which the whole website as an object based on the blocking algorithm and analysis of semantic characteristics, the experimental result shows high-accuracy level classification in extracted websites with good practicability.