Abstract: Objective To screen new serum metabolic biomarkers for different drug resistance profiles of pulmonary tuberculosis (TB) and explore their mechanisms and functions. Methods We collected serum samples from TB patients with drug sensitivity (DS), monoresistance to isoniazid (MR-INH), monoresistance to rifampin (MR-RFP), multidrug resistance (MDR), and polyresistance (PR). The metabolites in the serum samples were extracted by oscillatory and deproteinization for LC-MS/MS analysis, and the results were normalized by Pareto-scaling method and analyzed using Metaboanalyst 4.0 software to identify the differential metabolites. The differential metabolites were characterized by function enrichment and co-expression analysis to explore their function and possible pathological mechanisms. Results Compared with the DS group, 286 abnormally expressed metabolites were identified in MR-INH group, 362 in MR-RPF group, 277 in MDR group and 1208 in PR group by LC-MS/MS analysis. Acetylagmatine (P<0.05), aminopentol (P<0.05), and tetracosanyl oleate (P<0.05) in MR-INH group; Ala His Pro Thr (P<0.001) and glycinoprenol-9 (P<0.05) in MR-RFP group; trimethylamine (P<0.05), penaresidin A (P<0.05), and verazine (P<0.05) in MDR group; and PIP (18:1(11Z)/18:3(6Z,9Z,12Z)) (P<0.001), Pro Arg Trp Tyr (P<0.001), N-methyldioctylamine (P<0.001), and phytolaccoside E (P< 0.05) in PR group all showed significant differential expressions. Significant differential expressions of phthalic acid mono-2-ethylhexyl ester (P<0.05) and eicosanoyl-EA (P<0.05) were found in all the drug resistant groups as compared with DS group. Conclusion Acetylagmatine, aminopentol, tetracosanyl oleate, Ala His Pro Thr, glycinoprenol-9, trimethylamine, penaresidin A, verazine, PIP(18:1(11Z)/18:3(6Z,9Z,12Z)), Pro Arg Trp Tyr, N-methyldioctylamine, phytolaccoside E, phthalic acid mono-2-ethylhexyl ester, and eicosanoyl-EA are potentially new biomarkers that indicate monoresistance, multi-drug resistance and polyresistance of Mycobacterium tuberculosis. The combined use of these biomarkers potentially allows for assessment of drug resistance in TB and enhances the diagnostic sensitivity and specificity.