This paper presents a novel approach to detect SYN flood Distributed Denial of Service (DDoS) attacks using the ResNet 50 architecture. DDoS attacks, known for their ability to disrupt normal network traffic through overwhelming floods, have evolved in complexity, rendering traditional detection methods inadequate. Our methodology is threefold: data acquisition from both simulated and real-world environments, data processing where network traffic data is converted into 2D images, and attack detection using a Convolutional Neural Network model. The model’s performance was rigorously evaluated, demonstrating exceptional accuracy of 97.5%, which indicates the model’s effectiveness in controlled environments. The ResNet-50 based model shows promising results in accurately classifying network traffic and identifying DDoS attacks. This not only validates the effectiveness of deep learning in cybersecurity but also opens avenues for more robust and adaptable network defense mechanisms.