Abstract: DDoS attack will cause devastating damage to SDN,so it is particularly important to detect DDoS attack efficiently and accurately.To solve this problem, a detection plan for DDoS attack based on Rényi-RF-XGBoost in SDN environment is proposed.Rényi entropy is used to extract features and improve the random forest, then integrated with XGBoost through ensemble learning to classify and predict network traffic, thus realizing the detection of DDoS attacks.In addition, cross entropy loss and out of bag error are used to evaluate the proposed model, and the experimental results are validated by real-time observation through the relevant testing metrics.The results show that the proposed method not only has lower cross entropy loss and out-of-bag error, but also improves the detection precision, accuracy rate and recall rate, shorts the detection time and reduces the false alarm rate compared with other methods.