Abstarct: Fast-Flux botnets are becoming a type of cyber threat that can seriously endanger the Internet and online software by compromising the security of online systems and services. These botnets use a technique called fast-flux, which makes it very difficult to detect infected systems. To address this challenge, this thesis provides a comprehensive study of fast-flux botnets, including their aspect, behavior, and impact on the Internet. It then goes on to propose a new place to identify fast-flux botnets by analyzing and analyzing Domain Name System (DNS) traffic. This research addresses the problem of fast flux botnet detection using a novel approach baxsed on Convolutional Neural Networks (CNNs). The effectiveness of the proposed method is evaluated through experimental experiments using real-world fast-flux botnet datasets. Evaluation results can be quickly compared with advanced techniques for data recognition. The results show that the proposed method performs better than the existing techniques in terms of accuracy, efficiency and robustness. Also, the proposed method is able to identify botnets with fast flux in real time and with high accuracy, which will be a valuable contribution to the field of network security. Overall, this thesis presents a solution to detect fast flux botnets using insight algorithms and provides valuable functions for the behavior and impact of fast flux botnets on the Internet. It helps network security by providing a practical solution and diagnosis of this persistent crisis.