Peony root is an important herbal drug used as an antispasmodic analgesic. To evaluate peony roots with different botanical origin, producing areas, and post-harvest processing, 1 H NMR-based metabolomics analysis was employed. Five types of monoterpenoids, including albi orin (4), paeoni orin (6), and sulfonated paeoni orin (25), and six other compounds, including 1,2,3,4,6-penta-O-galloyl-β-d-glucose (18), benzoic acid (21), gallic acid (22), and sucrose (26) were detected in the extracts of peony root samples. Among them, compounds 4, 6, 18, and total monoterpenoids including 21 were quanti ed by quantitative 1 H NMR (qHNMR). Compound 25 was detected in 1 H NMR spectra of sulfurfumigated white peony root (WPR) extracts indicating that 1 H NMR was a fast and effective method for identifying sulfur-fumigated WPR. The content of 26, the main factor affecting extract yield, increased signi cantly in peony root after low-temperature storage for one month, whereas that in WPR did not increase due to the boiling treatment after harvesting. We investigated the impact of preprocessing methods to such analysis for NMR data from commercial samples, resulting that the data matrix transformed from qHNMR spectra and normalized to internal standard were optimum for multivariate analysis. The multivariate analysis demonstrated that among commercial samples derived from P. lacti ora, peony root samples in Japanese market (PR) had high contents of 18 and 22, and red peony root (RPR) samples had high content of monoterpenoids represented by 6; and among RPR samples, those derived from P. veitchii showed higher contents of 18 and 22 than those from P. lacti ora. The 1 H NMR-based metabolomics method coupled with qHNMR was useful for evaluation of peony root and would be applicable for other crude drugs.Both HPLC and LC-MS are targeted analysis methods. To focus on comprehensive constituents of peony root, 1 H nuclear magnetic resonance( 1 H NMR)-based metabolite pro ling method, known as untargeted method, was employed to investigate the component variations caused by multiple factors (differences in origins, producing areas, and post-harvest processing). Previous studies showed that 1 H NMR pro ling offers analytical advantages over HPLC and LC-MS: uncomplicated sample preparation, comprehensive detection of hydrogen-containing compounds, and time savings [14]. However, some points were not clear regarding the preprocessing method of NMR data for multivariate analysis. For the measurement of NMR spectra as the basis of the data matrix in multivariate analysis, both 1 H NMR [15][16][17] and quantitative 1 H NMR (qHNMR) [18][19][20] have been used. Additionally, the transformation from original NMR spectra data to the data matrix used for multivariate analysis needs to be normalized to minimize the in uence of experiment and instrumental variables [21]. There are two normalization methods: 1) normalization to the peak area of the internal standard [15,17,22], 2) normalization to total intensity [16, 18,23]. However, the method m...