In this thesis, swarm airports are introduced as a way of obtaining continuous operation in an autonomous drone swarm. An airport acts as a service station where the drones can replenish their energy autonomously. When incorporating several airports in the same system, one can operate a large drone swarm without the need for human interaction. This simplifies the use of drone swarms in applications such as surveillance and delivery. As there is little or no previous research on swarm airports, there are many issues that could have been addressed. In this thesis, the focus is on how one can optimize the airport selection when a drone requires a battery change. The thesis proposes a possible solution using a method inspired by response threshold, from the field of swarm intelligence. The advantages of using this method lie in the decentralized control, and ability to operate with limited communication. In order to perform experiments with the proposed method, a simulation tool was created. The simulation program provides a framework and visualization tool in order to better understand the system, as well as making it easier to compare methods with the same configurations. In order to understand how well the proposed method works, it is compared to benchmark methods inspired by mathematical optimization and random decision making. The main experiments presented in this thesis demonstrate how the different methods perform in terms of active drones in a system consisting of four swarm airports. The results of the swarm method proved adequate when the airports are spread out to several locations, but the method still requires some improvement in order to achieve the same results as the benchmark methods. When changing the configurations such that all airports are at the same location, the method performs well compared to the benchmark methods, indicating that using swarm optimization can be favorable when solving airport selection problems.