<p>In this paper<b>,</b> Coati Optimization Algorithm-optimized deep feedforward neural network fused long short-term memory (DFNN-LSTM) model is proposed and implemented for thwarting DDoS attacks in software defined network (SDN)-cloud computing environment. In this DFNN-LSTM approach, both the control plane and data plane is effectively used for defending DDoS attacks in software defined network (SDN)-cloud computing environment. This deep learning used the benefits of deep feedforward neural network fused long short-term memory (DFNN-LSTM) for extracting new and potential features from the statistics derived from the data traffic at the control plane. The significant features that are extracted by the adopted deep learning model from the data traffic statistics include type of service header, transport layer protocol header, inter-arrival time of packets and unknown IP destination address. Then the mean arrival bit rate of the switches with the unknown destination address if determined at the data plane such that mean number of flows and IP options header can be identified for facilitating the detection of DDoS attacks in the SDN network. It estimated the trust value and updates it periodically such that suspicious senders are blocked whenever the deep learning model detects the attack. The experimentation is conducted with the dataset which is generated using the feature extractions and computations determined over the normal and attack packets through the inclusion of a classifier. The performance metrics used for evaluating this proposed research work can be achieved using accuracy, false alarm rate, detection rate and precision.</p>