The invention relates to an audio/video intelligent catalog information acquisition method facing to a wide area network, belonging to the field of computer application, which is characterized in that a weighting algorithm based on position factors of keyword characteristic items is offered; different weighting factors are endowed to the characteristics of different positions in a file so that the theme similarity of webpage contents can be more accurately calculated; a link with higher theme similarity is optimally selected by comprehensively utilizing three aspects of the factors of the similarity of the webpage contents, the URL (Uniform Resource Locator) catalog level information of an ultra-link and the anchor text information of the ultra-link; the cataloguing information of the searched theme webpage is automatically extracted by adopting an information extraction method based on the body and HTML (Hypertext Markup Language); and the extracted cataloguing information is standardized by adopting an improved semantic similarity calculation method. The invention has the advantages that the description item information can be intelligently and automatically provided to a lister; the labor capacity of workers is reduced; the cataloguing efficiency is enhanced; and the method can adapt to different requirements of professional and non-professional listers and wide area network environment.