Software Defined Networking in Cloud paradigm is most suitable for dynamic functionality and reduces the computation complexity. The routers and switches located at the network's boundaries are managed by software-defined netwrking (SDN) using open protocols and specialised open programmable interfaces. But the security threats often degrade the performance of SDN due to its constraints of resource usage. The most sensitive components which are vulnerable to DDoS attacks are controller and control plane bandwidth. The existing conventional classification algorithms lacks in detection of new or unknown traffic packets which are malicious and results in degradation of SDN performance in cloud resources. Hence, in this paper double filtering methodology is devised to detect both known and unknown pattern of malicious packets which affects the bandwidth of the control panel and the controller. The case-based reasoning is adapted for determining the known incoming traffic patterns before entering the SDN system. It classifies the packets are normal or abnormal based on the previous information gathered. The traffic patterns which is not matched from the previous patterns is treated as indeterministic packet and it is defined more precisely using the triplet representation of Neutrosophic intelligent system. The grade of belongingness, non-belongingness and indeterminacyis used as the main factors to detect the new pattern of attacking packets more effectively. From the experimental outcomes it is proved that DDoS attack detection in SDN based cloud environment is improved by adopting CBR-NIS compared to the existing classification model.