Background: Gemcitabine-based chemotherapy is the standard first-line treatment for advanced cholangiocarcinoma (CCA). However, drug resistance frequently develops, leading to poor prognosis. The metabolic alterations underlying gemcitabine resistance in CCA remain poorly understood. This study aimed to identify plasma metabolic signatures associated with gemcitabine resistance using untargeted metabolomics. Methods: A total of 52 advanced CCA patients treated with gemcitabine plus cisplatin were enrolled. After 3 months of treatment, patients were classified according to RECIST criteria as responders (partial response or stable disease, n=26) and non-responders (progressive disease, n=26). Pre-treatment plasma samples were collected and analyzed by UHPLC-MS/MS-based untargeted metabolomics. Multivariate statistical analysis (OPLS-DA) and pathway enrichment were performed. Results: A total of 143 differential metabolites were identified (VIP>1.5, p<0.05). Bile acid metabolism, glycerophospholipid metabolism, and tryptophan metabolism were the most significantly enriched pathways. Compared with responders, non-responders showed significantly elevated levels of conjugated bile acids including taurocholic acid (TCA) and glycodeoxycholic acid (GDCA) (both p<0.01), as well as increased lysophosphatidylcholines (LPCs) (p<0.05). Additionally, the kynurenine/tryptophan ratio was markedly higher in resistant patients (p<0.01), suggesting enhanced IDO activity. A combined panel of three metabolites (TCA, LPC 18:2, and kynurenine) predicted gemcitabine resistance with an area under the curve (AUC) of 0.91 (95% CI: 0.83–0.97). Conclusions: Dysregulated bile acid, lipid, and tryptophan metabolism are closely associated with gemcitabine resistance in cholangiocarcinoma. These metabolic signatures may serve as predictive biomarkers for chemotherapy response and offer potential therapeutic targets to overcome resistance.