Objective: Metabonomics may identify potential metabolite markers for ovarian cancer diagnosis. The authors report a metabonomics study of ovarian cancer serum, aiming to identify a distinct serum metabolome of ovarian cancer with diagnostic potential. Materials and Methods: Serum metabolites from diagnosed ovarian cancer patients and healthy subjects were profiled using gas chromatography coupled with mass spectrometry (GC-MS). Results: Differential metabolites of 36 chemicals were identified with statistical tests of orthogonal partial least-squares-discriminant analysis (VIP > 1.2) and the t-test (p < 0.05), being able to differentiate ovarian cancer patients from the healthy controls. The authors observed an altered metabolome in ovarian cancer patients, including glycolysis and tricarboxylic acid (TCA) cycle, urea cycle, glutamine, fatty acids, and proline metabolism. A panel of three serum metabolite markers, containing beta-alanine, palmitic acid and proline, can differentiate ovarian cancer patients from healthy controls with an area under curve (AUC) of 0.995 in the receiver operating characteristic (ROC) analysis. Conclusion: The present results demonstrated that metabonomics is of great potential for finding a non-invasive diagnostic method for the detection of ovarian cancer.