The rapid proliferation of 5G technology, with its high-speed connectivity and extensive device interconnectivity, has created a broader attack surface for such cyber threats. Advanced detection techniques are necessary since current solutions, which aim to counter DDoS attacks, frequently fall short according to accuracy and efficiency. Using the CICDDoS2019 dataset, this study assesses how well AI-driven algorithms identify and mitigate DDoS attacks. The methodology includes data preprocessing, feature selection with an extra tree classifier, and implementation LSTM model and comparison with Decision tree and Neural network. Impressively outperforming DT and NN, the LSTM model achieved 99.9% accuracy, 99% precision, recall, and F1-score. Under different attack intensities, experimental setups analyzed the bandwidth and latency impacts of simulated DDoS attacks in a 5G network slicing context. A result prove LSTM’s ability and suitability in handling sequential data, which makes it ideal for DDoS detection in dynamic 5G networks. There are plans for future work that focus on rectifying the overfitting issues in LSTM, introducing more complex attack scenarios, and further optimizing the LSTM solution for scalable real-time applications in 5G networks.