In today’s interconnected internet landscape, monitoring for misuse by malicious users is crucial. DDoS attacks are a highly prevalent malicious technique; it is a transaction that attempts to flood a server’s normal traffic, network, or services. The last few years have witnessed a decrease in network availability with higher detection rates and complexity. Recently, ML-based detection methods have emerged as one of the most prominent approaches. This paper introduces a Multilayer Perceptron and Convolutional Neural Network (MLP-CNN) for classifying and predicting DDoS attacks. By injecting various DDoS attacks into regular data, an extensive dataset is compiled. The proposed model combines CNN with MLP, and its performance is assessed using a confusion matrix. Experimental results demonstrate the model’s efficiency with 97% accuracy, 96% precision, and 96.5% recall. The detection time for 41 attributes is 2 seconds, and for 9 attributes, it is 0.2 seconds.