The invention provides a method and a system for searching similar web pages based on DOM web page pruning, which comprises the following steps: formatting input data; obtaining a digital identification of the web page; Similar web page lookup: Looking for similar web pages in a database that stores digital identifiers for web pages. The method comprises the following steps: formatting the web page data: obtaining the HTML source code of the input web page; Web page purification: remove redundant information from the web page; Generating DOM tree of web page: DOM parsing the web page to generate DOM tree; Refined DOM tree: Traverse the DOM tree, find the smallest format node, delete the descendant nodes of the smallest format node, and form a new DOM tree; Text sequence conversion: a new DOM tree is traversed in depth first to get a text sequence of HTML tags. Digital identification generation: using simhash algorithm to calculate the text sequence to obtain the digital identificationof the web page. The method of the invention can improve the searching efficiency of the similar webpages in the massive webpages.