The rapid, accurate identification of antibiotic-resistant bacteria is crucial for effective clinical management and infection control. In this study, we elucidated a highly sensitive and reproducible surface-enhanced Raman scattering (SERS)-based microdroplet sensor for discriminating methicillin-resistant Staphylococcus aureus (MRSA) from methicillin-sensitive Staphylococcus aureus (MSSA). By engineering oligonucleotide-based core-satellite gold nanoparticles (AuNPs) as SERS nanotags and integrating them with aptamer-functionalized magnetic beads, the platform enables the specific capture and detection of target bacteria. The use of a droplet-based microfluidic system facilitates ensemble-averaged Raman measurements across >100 droplets, which significantly enhances the detection reproducibility, compared to that of conventional methods. The sensor enables quantitative detection of MRSA from mixtures containing both MRSA and MSSA by harnessing the selective binding properties of distinct MRSA SERS nanotags. This approach achieves low limits of detection and high selectivity, which addresses major challenges in SERS-based bacterial diagnostics. The novel microdroplet sensor offers a promising platform for rapid, accurate identification of antibiotic-resistant pathogens, with potential for broad application in clinical diagnostics and infectious disease management.