Pretreatments of chromatographic fingerprints are important for quality control of herbal medicines and they include data correction and data transformation. The data correction can reduce the variations of experimental procedures, and data transformation can put different weights on the different parts of the fingerprints. In this paper, a new target peak alignment (TPA) procedure has been proposed to correct the retention time shifts, multiplicative scattering correction (MSC) has been introduced for response correction. Then the similarity of the fingerprints with mean and median fingerprints is used to evaluate the quality of herbal medicines (HMs). Furthermore, different data transformation methods with kernel principal component analysis (PCA) have been applied to the data and their effects were discussed. The proposed approaches have been demonstrated by the essential oils data set of a herbal medicine, named Houttuynia cordata (HC), containing samples from different geographic origins. The experimental results indicate that the proposed approaches may be helpful in the quality control of herbal medicines by fingerprints.