The study examined plasma exosomes from patients with advanced lung cancer to identify biomarkers predictive of brain metastases. They extracted and sequenced total RNA from exosomes and analyzed it using BLAST software with GO and KEGG databases. The results showed a significant increase in exosomal miR-223-3p, miR-27a-3p and miR-27b-3p in patients with brain metastases. Concentrations of short-chain fatty acids: isobutyric (IBA), valeric (VA), isovaleric (IVA), and acetic (AA) also increased. Spearman correlation analysis confirmed a significant association between miR-27a-3p, miR-27b-3p and fatty acids IVA and VA. The combination of these exosomal miRNAs and metabolites in a multi-biomarker model achieved a high diagnostic performance with an AUC of 0.927. These biomarkers may serve to predict brain metastases in patients with advanced lung cancer.