The security measures of the airspace have become vital due to the significant growth of Unmanned Aerial Vehicles (UAVs). For countermeasure the intruder present in airspace, a group of UAVs, i.e., a UAV swarm can be preferred. There are several advantages of UAV swarm which help in safeguarding the danger area from the enemy. This research work offers a revolutionary technique that explores UAV network topology for the identification of intruder drones by merging Deep Learning (DL) and Metaheuristic methodologies. This work highlights the adaptable behavior of a swarm of UAVs toward finding the intruder by incorporating the asset of neural networks with metaheuristic optimization techniques. In this work, a novel framework is presented based on conventional and real-time data. Toward the development of UAV technology, the framework enhances the security and reliability of airspace missions. There are several challenges present while considering the proposed framework, which are discussed here. Future scopes are also presented in this work.