Metabolic characteristics of 39 human brain tumor tissues, including 15 astrocytomas, 13 fibroblastic meningiomas and 11 transitional meningiomas from 39 individual patients, have been studied using high resolution magic-angle spinning (HRMAS) ¹H NMR spectroscopy in conjunction with principal component analysis (PCA). With rich metabolite information, ¹H NMR spectra showed that the tumor-tissue metabonome was dominated by lipids, lactate, myo-inositol, creatine, choline metabolites such as choline, phosphocholine and glycerophosphocholine, amino acids such as alanine, glutamate, glutamine, taurine, N-acetyl-aspartate and glutathione. PCA of the tumor NMR spectra clearly showed metabonomic differences between low-grade astrocytomas and meningiomas whereas such differences were more moderate between fibroblastic and transitional meningiomas. Compared with meningiomas, the low-grade astrocytomas had higher levels of glycerophosphocholine, phosphocholine, myo-inositol and creatine but lower levels of alanine, glutamate, glutamine, glutathione and taurine. The N-acetyl-aspartate level was low but detectable in low-grade astrocytomas whereas it was not detectable in meningiomas. It is concluded that tissue metabonomics technology consisting of HRMAS ¹H NMR spectroscopy and multivariate data analysis (MVDA) offers a useful tool (1) for distinguishing different types of brain tumors, (2) for providing the metabolic information for human brain tumors, which are potentially useful for understanding biochemistry of tumor progression.