Low-altitude UAV swarms, characterized by low cost, easy deployment, and cooperative penetration, are progressively degrading the early-warning and response effectiveness of low-altitude defense systems. Radar, electro-optical and radio-frequency surveillance may have detection blind spots in scenarios involving low-altitude masking, complex electromagnetic environments and radio-silent flight, whereas acoustic early warning provides a passive, covert and flexibly deployable complement. Coordinated multi-UAV flight causes harmonic overlap, sound-field mixing and accumulated localization errors, which complicate source separation, direction finding and situation generation. This paper synthesizes research progress in acoustic signatures, system architectures, array front ends, robust recognition, multi-target sound source separation and localization, distributed networking and multimodal fusion. It further outlines a distributed integrated acoustic-optical-electronic-magnetic early-warning architecture, providing support for the development of an all-domain, three-dimensional collaborative early-warning technology system for low-altitude UAV swarms.