Dragon fruit has appealed much concern from consumers as a novelty fruit with potent nutritional and medicinal benefits. Dragon fruit quality warrants comprehensive evaluation, based on the contents of pigments and health-promoting natural compounds in different varieties. This study was aimed to evaluate the differences among dragon fruit varieties extracted with methanol–water (CD₃OD-D₂O) and methanol (CD₃OD) by proton nuclear magnetic resonance (¹H NMR)-based metabolomics approach. The variation features of the metabolite profiles were studied between varieties and origins of dragon fruit, considering the differences in principal component analysis (PCA). The hierarchical clustering analysis (HCA) based on score values of PCA model was also performed to analyze the distance between samples based on metabolites contents. The results of ¹H NMR spectra showed that the CD₃OD-D₂O extracts quantitatively differ from CD₃OD ones. In dragon fruit extracts, 36 metabolites were identified. The results demonstrated that the methanol and methanol/water extracted similar compounds with higher intensity in methanol. The metabolic differences among varieties were also shown for CD₃OD extracts by comparing both Pareto and UV scaling methods. The big size red fleshed dragon fruit (samples 2 and 3), growing in Israel were clustered similar to that growing in Thailand with the abundance of phenolic compounds. Glucose and fructose were more prominent in the yellow and white fleshed fruit (samples 4 and 5) growing in Israel. To support the obtained results two dimensional ¹H–¹H J-resolved and UHPLC-MS measurements were carried out. This research gain novel insights into the field as the first NMR metabolites fingerprinting of the major dragon fruit varieties. The correlations between DPPH, CUPRAC, antioxidant and metabolomic properties were also evaluated. The chemical markers associated with varieties of dragon fruit quality and their appearances were identified and can be utilized for the
basis of authentication purpose of this fruit.