UAVs that function as base stations (UAV-BS) are valuable supports when terrestrial infrastructure cannot fulfill service requirements. To facilitate effective UAV-BS deployment in fluctuating and energy-limited environments, this paper proposes a rotating shift (RS) deployment strategy alongside a deployment fine-tuning algorithm. The RS strategy provides continuous service by methodically alternating UAV-BS assignments, and the fine-tuning algorithm leverages distributional reinforcement learning to refine deployment choices. Simulations confirm the approach’s efficacy, showcasing improved energy efficiency and reliability over current methods.