This study evaluated the effectiveness of spectral reflectance techniques using a portable instrument combined with chemometrics to trace the geographical and mono e multifloral honey from Ortigueira, Brazil, a region with Protected Designation of Origin (PDO). Honey samples were collected from Ortigueira and 22 other locations in Paraná. Principal Component Analysis (PCA) revealed a trend of separation between the samples, though not entirely precise. Partial Least Squares Discriminant Analysis (PLS-DA) and Linear Discriminant Analysis (LDA) models were developed to enhance accuracy. PLS-DA demonstrated high geographical discrimination capability, with 97.1 % accuracy, while LDA performed well in training but had lower testing accuracy. For discriminating floral origins, PLS-DA achieved an accuracy of 82.4 %, highlighting the complexity of multifloral honey. Overall, PLS-DA outperformed LDA in distinguishing both geographic and floral origins. This study demonstrates the potential of spectral reflectance and chemometrics, particularly for the geographic discrimination of honey. Although PLS-DA was effective, improvements are needed for more accurate floral discrimination. The approach shows promise in combatting honey fraud and monitoring certified products, demonstrating how portable technologies can facilitate rapid, real-time decisions in food authentication and quality assessment.