This paper explores how machine learning approaches can be applied to detect and mitigate Distributed Denial of Service (DDoS) attacks within cloud-based systems. As the adoption of cloud computing continues to grow, safeguarding infrastructure from both large-scale and sporadic DDoS incidents has become increasingly important. The discussion focuses on the use of various supervised and unsupervised learning algorithms—such as Random Forest, Support Vector Machine (SVM), and Long Short-Term Memory (LSTM)—and how these techniques integrate with cloud-native security tools including AWS WAF, Shield, and Lambda. Publicly available datasets like CICDDoS2019 are analyzed to highlight the strengths and weaknesses of different models. In addition, the paper reviews recent studies emphasizing automated defense mechanisms and strategies for effective cloud deployment. It also considers how factors such as detection delay and mitigation procedures influence overall service reliability through real-world attack testing. By presenting these findings, the study aims to guide future researchers in developing secure and scalable architectures for cloud computing environments.