The ubiquitous usage of Internet of Medical Things (IoMT) enables real-time health monitoring in daily life. However, a huge amount of generated medical data causes great trouble to IoMT devices, owing to their power-constrained and computation-limited features. On the other hand, unmanned aerial vehicles (UAVs) can achieve a seamless wireless connection to IoMT devices. In this article, we leverage UAVs as the aerial fog nodes to provide computational resources to the IoMT devices. We aim to minimize the number of UAVs for covering the IoMT devices on the ground. This problem is modeled as a discrete nonlinear combinatory optimization problem with NP-hardness. A particle swarm optimization (PSO)-based strategy is put forward to solve this problem. In particular, we introduce greedy heuristic to speed up the searching process. The experimental results show that the proposed algorithms GASM_a and GASM_b can achieve better performance compared to the greedy and random approaches in terms of optimal values. In addition, the enhanced algorithm GASM_b can obtain the better result than GASM_a. Meanwhile, the execution time is also averagely increased by 16.51% compared to GASM_a. Increasingly, low hardware prices enable large-scale deployment of UAVs for IoMT in healthcare systems. Our study shows the advantages in wireless coverage for ground patients by UAVs in terms of efficiency, and cost that will contribute to the development of UAV assisted healthcare systems.