Small extracellular vesicles (EVs), including exosomes, in body fluids have important applications in the noninvasive liquid biopsy-based diagnosis of cancer. Current EV-based diagnostic techniques still face practical challenges, such as inefficient EV isolation. Here, we report an efficient, resource-free pre-enrichment approach using (-)-epigallocatechin-3-gallate (EGCG), a polyphenolic biomolecule, to isolate and detect exosomal microRNAs (miRNAs) in human blood plasma samples. Our system comprises three steps: (1) EGCG-mediated EV aggregation, (2) filter-based EV isolation, and (3) molecular beacon-based detection of target miRNA in EVs. Using blood samples from cancer patients with gastric cancer or hepatocellular carcinoma, we constructed an EGCG-assisted miRNA diagnostic system. For both cancers, the levels of target miRNAs (miR-21, -27a, and -375) in EVs were strongly correlated with those in the publicly available GEO database. Our approach, an easy-to-use method for efficient EV isolation and the detection of miRNA in clinical samples, is applicable for molecular diagnostics in precision medicine.