Recently, there are many denial-of-service (DoS) attacks by computer viruses or botnet. DoS attacks to web services are called HTTP-GET flood attack and threats of them increase day by day. In this type of attacks, malicious clients send a large number of HTTP-GET requests to the target web server automatically. Since these HTTP-GET requests have legitimate formats and are sent via normal TCP connections, an intrusion detection system (IDS) can not detect them. In this paper, we propose HTTP-GET flood detection techniques based on analysis of page access behavior.We propose two detection algorithms, one is focusing on a browsing order of pages and the other is focusing on a correlation with browsing time to page information size. We implement detection techniques and evaluate attack detection rates, i.e., false positive and false negative. The results show that our techniques can detect the HTTP-GET flood attack effectively.