Abstract: Since the independent design between web contents and styles of National Level Excellent Courses,web page segmentation algorithm can hardly run．We present a rule-based preprocessing algorithm of web page segmentation to create correlation between tags and style information．The algorithm consists of three steps:first,get the style information；second,associate styles with tags；third,output HTML and PerfectNode which is associated class list．We selected 100pages from the National Level Excellent Courses randomly to run the preprocessing algorithm．Experimental results show that the algorithm can associate tags with styles efficiently,which can solve the problems that web page segmentation algorithm cannot run.

Key words: Web page segmentation,Preprocessing algorithm,Cascading style sheets,Style information