In this paper, we propose a three-dimensional path planning algorithm for a large-scale swarm. The proposed algorithm isbased on the Eulerian approach that controls spatial characteristics which is an agent density vector. This algorithm has three phases,which are environment abstraction, optimal transport, and collision avoidance, respectively. The algorithm can be applied to a largescaleswarm of 1000 agents or more, and it guarantees collision avoidance between obstacles existing in the given flight environmentand the surrounding agents. We carried out simulation using two scenarios, and as a result, it was confirmed that the proposedalgorithm is suitable for three-dimensional path planning for a large-scale swarm.