Rheumatoid arthritis (RA) is the most severe type of chronic inflammatory disease and has always been a research hotspot in different fields. In this study, a non-targeted metabonomics approach was carried out to profile metabolic characteristics of RA and its Chinese medicine subtypes by using liquid chromatography-mass spectrometry (LC-MS) and gas chromatographymass spectrometry (GC-MS). Plasma samples of 57 RA patients and 23 healthy controls were collected. On the basis of the traditional Chinese medicine (TCM), RA patients were classified into two main patterns, the cold pattern and the heat pattern. By using univariate and multivariate data analysis, we found that the RA patients presented diverse dysfunctions in inositol phosphate metabolism, lipid metabolism, amino acid metabolism, glucose metabolism, ascorbate metabolism, glyoxylate and dicarboxylate metabolism. The metabolic phenotypes were different between the RA cold pattern and the RA heat pattern. Compared with the RA cold pattern, the RA heat pattern showed elevated plasma concentrations of glycochenodeoxycholate, proline, saturated and mono-unsaturated phosphatidylcholine (PC) but decreased levels of urea, free fatty acid (FFA) and polyunsaturated PC. Our data show that metabonomics is a valuable tool in disease and TCM subtype research.