Page Rank algorithm is the most classic web structure mining algorithm,but it exists a problem of topic drift that make the webpage irrelevant to the query subject in search results. Based on the analysis of the pagerank algorithm, it proposes using the euclidean distance to compute the topic similarity and integrate into the traditional pagerank algorithm forming an improved webpage ranking algorithm. And this algorithm is applied to the cloud computing enviroment to explore the process of pagerank algorithm in the mapreduce programming model.