In this study, we designed an emergency navigation system for swarm drones operating without Global Navigation Satellite System (GNSS) support. The system is intended to facilitate emergency landings using Visual-Inertial Odometry (VIO) based on mission cameras when LiDAR-based Simultaneous Localization and Mapping (SLAM) is deemed unreliable. To reduce computational load on the mission computer, LiDAR is utilized for position estimation during regular mission operations, while the mission camera is employed for position estimation only in emergency situations. For application in swarm drones, emergency landing zones are further subdivided to prevent collisions among drones. To validate the emergency navigation system, various scenarios were created and simulated using the Gazebo simulator. The results confirmed that, during indoor swarm drone operations, the proposed emergency navigation system allows safe and collision-free landings in emergency situations.