Background Osteosarcopenia, defined by the coexistence of sarcopenia and osteoporosis, is increasingly recognized as a high-risk geriatric musculoskeletal syndrome. Although it is commonly diagnosed by applying separate criteria for sarcopenia and osteoporosis, whether osteosarcopenia has a distinct systemic metabolic phenotype remains unclear. This study aimed to characterize the serum metabolomic profile of osteosarcopenia and to identify discovery-level candidate metabolites that may distinguish osteosarcopenia from osteoporosis alone. Methods We conducted a cross-sectional untargeted metabolomics study using high-resolution liquid chromatography–mass spectrometry in 104 older adults, including healthy controls (HC, n = 30), participants with osteoporosis only (OP, n = 30), sarcopenia only (IS, n = 14), and osteosarcopenia (OS, n = 30). Because the IS group was relatively small, comparisons involving IS were treated as exploratory and hypothesis-generating. The primary mechanistic and biomarker analyses focused on the OP versus OS comparison. Differential metabolites, enriched pathways, and preliminary discriminatory performance were assessed using multivariate modeling, false discovery rate correction, pathway enrichment analysis, covariate-adjusted regression, and receiver operating characteristic analysis. Age, sex, body mass index, and estimated glomerular filtration rate were included as covariates. Results Osteosarcopenia showed a distinct serum metabolomic profile, particularly when compared with osteoporosis alone. In the OP versus OS comparison, differential metabolites were mainly involved in amino acid metabolism, energy metabolism, purine metabolism, and membrane lipid remodeling. Key alterations included reduced branched-chain amino acids, decreased hypoxanthine, increased creatine, and changes in sphingomyelin and phosphatidylcholine species. Pathway enrichment analysis highlighted valine, leucine, and isoleucine degradation; arginine and proline metabolism; and purine metabolism as significantly enriched pathways. Among the discovery-level candidate metabolites, hypoxanthine and leucine showed apparent discriminatory potential for distinguishing OS from OP, with area under the curve values of 0.87 and 0.79, respectively. These values should be interpreted as discovery-set estimates rather than externally validated diagnostic performance. Conclusions Osteosarcopenia was associated with a multi-pathway serum metabolic signature that differed from osteoporosis alone, supporting the possibility that osteosarcopenia represents an integrated musculoskeletal phenotype rather than merely the additive coexistence of muscle and bone loss. The identified metabolites provide preliminary mechanistic insight and discovery-level biomarker candidates, but they require targeted quantitative validation and external cohort confirmation before clinical application. Clinical trial registration http://www.chictr.org.cn , identifier ChiCTR2500113401.