Abstract <p>Atmospheric pressure laser plasma ionization technique is studied in application to the problem of tea classification according to the composition of volatile compounds without sample preparation. Tea leaf samples from nine different trade names were studied. The obtained mass-spectrometric data were clustered using chemometric methods without detailed component-by-component interpretation of the mass spectra. A comparative study of six methods, i.e., principal component analysis, K-means, and hierarchical clustering (unsupervised methods), as well as support vector machine, logistic regression, and multilayer perceptron neural network (supervised methods), was performed. It was shown that the most effective method for classifying tea samples is logistic regression combined with recursive feature elimination with cross-validation.</p>