The protein glycosylation of extracellular vesicles (EVs) is involved in cellular recognition and emerges as a promising biomarker for cancer diagnosis. However, the lack of efficient labeling and high-resolution detection strategies limits their clinical application. Herein, we developed a self-localized plasmonic nanocavity strategy to analyze the glycosylation characteristics of glioblastoma EVs. First, an engineered phospholipid bilayer structure with a Au nanoring array was designed to capture EVs and induce membrane fusion. Relying on the multifunctional proximity labeling process, a peroxidase-induced proximity labeling was designed to label sialic acid on programmed cell death ligand 1 (PD-L1) of EVs. Based on the identification and labeling process of EVs, the plasmonic nanocavity was self-localized with Au nanocubes and achieved the spontaneous location of MoSe