Gastric cardia adenocarcinoma (GCA) has a high mortality rate worldwide; however, current early diagnostic methods lack efficacy. Therefore, the aim of the present study was to identify potential biomarkers for the early diagnosis of GCA. Global metabolic profiles were obtained from plasma samples collected from 21 patients with GCA and 48 healthy controls using ultra-performance liquid chromatography/quadrupole-time-of-flight mass spectrometry. The orthogonal partial least squares discrimination analysis model was applied to distinguish patients with GCA from healthy controls and to identify potential biomarkers. Metabolic pathway analysis was performed using MetaboAnalyst (version 4.0) and revealed that 'glycerophospholipid metabolism', linoleic acid metabolism', 'fatty acid biosynthesis' and 'primary bile acid biosynthesis' were significantly associated with GCA. In addition, an early diagnostic model for GCA was established based on the relative levels of four key biomarkers, including phosphorylcholine, glycocholic acid, L-acetylcarnitine and arachidonic acid. The area under the receiver operating characteristic curve revealed that the diagnostic model had a sensitivity and specificity of 0.977 and 0.952, respectively. The present study demonstrated that metabolomics may aid the identification of the mechanisms underlying the pathogenesis of GCA. In addition, the proposed diagnostic method may serve as a promising approach for the early diagnosis of GCA.