The demand for future wireless communication systems is being satisfied for various circumstances through unmanned aerial vehicles (UAVs), which act as flying base stations (BSs). In this letter, we propose an ellipse clustering algorithm that maximizes the user coverage probability of UAV-BSs and avoids inter-cell interference with minimal transmit power. We obtain the coverage of each UAV by adjusting its antenna half-power beamwidth, orientation, and 3D location by minimizing the path loss of the cell-edge user. Simulation results confirm that the proposed algorithm achieves high system throughput and coverage probability with lower transmit power compared to conventional algorithms.