In the quest for a rapid and cost-effective tool for determining the botanical origin of monofloral honey, the analytical capabilities of coated blade spray mass spectrometry (CBS-MS) were investigated. To this aim, the chemical profiles of 64 honey samples from seven different botanical origins (acacia, dandelion, chestnut, rhododendron, citrus, sunflower, and linden) were captured by the absorbent of the coated blades and then analyzed by mass spectrometry. An exploratory analysis was performed by principal component analysis (PCA) to generate a graphical representation of the CBS-MS data that allows the discovery of patterns, outliers, or relations between types of honey in an unsupervised manner. Additionally, the performances of four different classification algorithms (least absolute shrinkage and selection operator (LASSO), random forest (RF), and neural network (NNET) partial least squares discriminant analysis (PLS-DA)) were built up and compared. The performances of the four classifiers were verified by a 50 times-repeated, 5-fold-cross-validation and permutation test. Although all classifiers performed well, the RF showed significantly higher performances in cross-validation (with area under the curve (AUC) of 0.99, overall accuracy 0.94, Kappa 0.93, sensitivity 0.94, and specificity 0.99). Moreover, the permutation tests showed the models were not overfitted. Finally, to determine the molecular identities of the ions that most contribute to the classification, extracts from the same honey samples were prepared, analyzed by liquid chromatography coupled to high resolution tandem mass spectrometry (LC-HRMS/MS), and the most significant features were annotated. This proof-of-principle work warrants a future large-scale study to validate and challenge this CBS-MS-based method with a greater number of honeys from different years and geographical origins.