With the spread of computer, the increase of malware is a serious problem. In some malware, damage caused by botnet is a serious problem. Botnets perform the attack by remote control. The purpose of the present work is to suppress the botnet activity by detecting the C&C traffic through well-suited observations. There already exists many detection techniques, most of which focus on IRC-based botnet, and very little focus on HTTP-based botnet, even less, which include comparisons between both detection techniques. In this work, we focus on the HTTP-based botnet, and in order to classify normal HTTP session and C&C session, we make use of Support Vector Machine.