Abstract: Automatic classification and authentication are increasing in importance in plant science, since there is a need in the agrofood industry to protect the product credibility, as well as to improve the online monitoring procedures. In this context, the use of spectral analysis to classify and authenticate agricultural products in a rapid and nondestructive way has been benefited by the evolution of electronics and computational methods, in particular the chemometrics. The spectral signatures of plants tend to vary with species, varieties, age, internal cellular structure, environmental conditions, chemical composition, and nutritional level, among other properties. When plant spectra present noise or do not visibly differ among themselves it is possible to apply preprocessing techniques, removing irrelevant information and improving the computing efficiency of the mathematical models. There are many chemometric methods which can be successfully used to classify or authenticate plants, each one with its own algorithm for determining how to discriminate at best the different groups. Thus, this review was conducted to summarize the most used chemometric methods in authentication and classification of plants based on spectral data, also presenting recent applications of these techniques.