Distributed Denial of Service (DDoS) attacks exploit cloud resources’ shared and expandable features, making them very hard to manage. To find DDoS attacks in cloud setups, traffic analysis tools watch how data flows, analyze patterns, look for strange patterns, and learn past actions. Progressively, analyzing traffic patterns on the network can help identify DDoS attacks. Deep Learning (DL) methods are gradually used to estimate and detect attack schemes through the analysis of historical data. This study aims to enhance the effectiveness and efficiency of anomaly detection models deployed in cloud computing architecture for combating DDoS attacks. This research focuses on DDoS attacks as the major threat encountered in cloud computing environments to propose an improved protection model from DDoS attacks. Subsequently, it proposes a Collaborative Long Short-Term Memory (CLSTM) model for securing cloud computing environments against DDoS flooding attacks. The proposed CLSTM model integrates DL and MAS algorithms to a distributed deep learning network to enhance DDoS attack detection in cloud computing environments. The model aims to detect attacks distributed across different network nodes and with different patterns. The CLSTM model is evaluated under key performance metrics in the testing phase, and it achieves an accuracy of 93.88%, a precision of 90.85%, a recall rate of 87.82%, and an F1 score of 88.83%.