Journal of Information Organization Pei Tang, Shanhong Zhu 7 4 2017 http://www.dline.info/jio/fulltext/v7n4/jiov7n4_3.pdf User behavior analysis is important for both Web information retrieval technologies and commercial search engine algorithms. With the expansion of information data, the current search engine is facing some serious problems, such as limited storage space and computing power. The paper discussed the shortcomings and technical bottlenecks of the current traditional search engine. Then, in the understanding of search engine features and technical requirement, it improved the system by means of the cloud computing architecture. With the combination of the static analysis of user behaviour and real-time monitoring, real-time acquisition of Web log and user to access the context information of the page, the paper tested the whole system performance in the laboratory environment, demonstrated the superiority of the system by analysis of experimental data.