In recent years, to protect source-location-privacy (SLP) in underwater acoustic sensor networks (UASNs), some schemes through the collaboration of multi-autonomous underwater vehicle (AUV) have been proposed. However, the long end-to-end delay in these schemes leads to untimely data delivery. To address this issue and enhance SLP protection, a low-delay SLP protection scheme with multi-AUV (LDSLP-MA) collaboration for UASNs is proposed in this article. In the LDSLP-MA scheme, a multipath technique including multipath routing as well as multi-AUV collaboration is employed to enhance SLP protection. Additionally, through strategically assigning dwelling and target areas for AUVs, the delay taken by multi-AUV scheduling is minimized while the diversity of data transmission paths and SLP protection is enhanced. Specifically, the optimal target area is selected through gray relational analysis. Simulation results demonstrate that the LDSLP-MA scheme achieves an extended safety period, decreased energy consumption, and reduced delay compared to other schemes. Notably, in comparison to multi-AUV collaboration-based SLP protection schemes like the push-based probabilistic method for SLP protection (PP-SLPP) and stratification-based SLP (SSLP), LDSLP-MA increases the safety period by over 100%, reduces delay by over 82%, and lowers average node energy consumption by over 65%.