Honey is a complex matrix that contains a wide range of compounds. This rich composition is influenced by diverse environmental factors, including geographic and botanical origin. Honey has been among the most commonly tampered foods worldwide, with improvements in techniques to do it. Accordingly, there is a recurring need for new techniques and methods to assess the honey's metabolic profiles to distinguish adulterated from non-tampered samples. In this sense, this study aimed to determine the chemical profiles of honey samples from the eleven agroecological zones of the Santa Catarina State (southern Brazil), collected in the 2019-2020 and 2020-2021 harvest seasons through 1D- and 2D-NMR. As a result, a series of metabolites was identified and their concentrations measured in samples. Further, the metabolomic dataset was used for building descriptive models through chemometric techniques, in order to discriminate honey samples according to their geographic and botanical origins and harvest season effect. Twenty-one metabolites were identified, with predominance of glucose and fructose in all samples. Two other carbohydrates (sucrose and maltose) were identified in lower concentrations, in addition to amino acids, organic acids, ketone, alcohol, ester, and alkaloids. No discrepant 1H NMR resonances that could indicate fraud were detected in the spectra. By PCA, it was possible to find clusters with similar geographic origins, i.e., agroecological zones, and botanical origins. In this regard, patterns of composition were detected for honey samples of Eucalyptus spp. and Hovenia dulcis species, which presented acetoin and kynurenate, respectively, in higher concentrations. Taking together, the results allowed demonstrating that NMR spectroscopy coupled to chemometrics is an effective experimental approach to characterize Brazilian honey regarding their geographic origin and season of collection, despite the huge floral diversity available in that country for bee forage.