Abstract: A growing number of botnets threats recently has grown to the level of world wide concerns. This dangerous phenomenon emerges drastically and offers undefined capability to attack the global internet security never seen before. As time evolves, the incremental numbers of botnets attack have been recorded with types of variants such as peer-to-peer (P2P) have been discovered. Attentively, botnets attack nowadays is typically declared as an advance malware due to its ability to smokescreen itself as a benign P2P application which make it difficult to detect and shut down and also easily to escape itself. Alarming on this crisis, many studies propose on detection, prevention and mitigation techniques as the precaution action. Hence, this study addresses in-depth review on a wild scale for botnets detection techniques. Technically, the survey classifies the detection techniques into five categories based on its anomaly, signature, DNS, data mining and hybrid technique. To enrich the level of understandings on the strategy, this study also highlights the importance of such characteristics as type of technique, approach, response time, type of botnets, detection parameter, metric and variants. In addition, this study offers detail discussion about botnets detection techniques which is beneficial for botnets investigation and helpful to other researches for immediate references.