By investigating 80 samples of ripe, unripe and processed rape honey collected from three regions (Anhui, Inner Mongolia and Jiangsu) in China, a method based on NOESY 1D combined with O2PLS-DA was developed to discriminate the mentioned three types honey. NOESY 1D spectra of representative honey samples were first compared, and some obvious differences among the three types of honey were observed. O2PLS-DA model was applied to discriminate the collected samples successfully. The segmental bin was selected based on the greatest R²Y and Q² values. Three types of honey from three regions could be successfully discriminated, although there was partially overlap among some of the unripe and processed honey samples, but they can be discriminated in 3D graph. The developed method could be used to identify ripe, unripe and processed honey from same region, all R²Y and Q² are higher than 0.97 and 0.81, respectively. This method could be applied to discriminate the ripe, unripe and processed rape honey from the same region. The advantages of this method are small sample amount, short time, good reproducibility and reliable results.

ripe honey, unripe honey, processed honey (concentrated honey), nuclear magnetic resonance Spectroscopy (NMR), orthogonal bidirectional projections to latent structures-discriminant analysis (O2PLS-DA)