An efficient, effective and reliable method for authentication of herbal medicines (HMs) based on a novel wavelet-enhanced automated three-dimensional chromatograms classification system. The method takes advantage of the character of 3D chromatograms which appear as functional curves and the preprocessing power of wavelet transform (WT) to minimize the interferences from noises and baseline drifting of the raw data, and at the same time significantly reduces the data sizes. This compressed data retain the essential features of the original raw data, whereby leading to efficient and reliable method for identifying or authenticating herbal medicines or other chemical mixtures of great complexity.