“Cold-hot nature” is the key part of the theory on herb properties in traditional Chinese medicine (TCM). Traditionally, the cold-hot natures of medicinal herbs were ascertained mainly dependent on medicinal experience. Nowadays systematic methods considering the material basis from a variety of chemical substances are needed and may be reliable to differentiate cold-hot nature of various herbs. In the present study, nuclear magnetic resonance spectroscopy of proton (1H NMR) was combined with pattern recognition techniques for cold-hot nature classification of Chinese medicinal herbs for the first time. 1H NMR determination was carried out on 29 kinds of Chinese medicinal herbs. Principal component analysis (PCA) and Fisher discriminant analysis (FDA) then were applied to analyze the large 1H-NMR spectra data set obtained. The results indicated that the combination of 1H-NMR and FDA may provide a more efficient means for the research and exploration of cold-hot nature of medicinal herbs.