A methodology that combines fluorescence spectroscopy, image polarimetry, and an analytical technique for processing experimental data to characterize and classify Mexican monofloral and multifloral honey is presented. Honey samples were excited with a 397 nm laser beam to produce a fluorescence signal, polarization images were acquired using a polarized CMOS sensor to perform polarimetric measurements, and a spectrometer was employed for measuring the produced fluorescence spectra. Image-based measurements are more appropriate for studying crystalized honey samples than point-signal measurements. Honey fluorescence produces partially polarized light when excited with linearly polarized light. According to our results, the degree of linear polarization is related to the botanical origin of the honey samples and their crystallization state. Three different classification methods are presented, and the results indicate a good agreement among them.