To solve the problem of automatic stylistic classification of texts, methods of discriminant analysis were applied. The trigram index (TI), the bigram index (BI), their ratio (TI/BI), the text compressibility index (Deflate), as well as information indexes: the ratio of order and chaos in the system – the so-called R-function (Rf), and the development function (Df) were considered as possible classification factors. Optimal combinations of indexes for solving this problem are found. It was possible to achieve a significant improvement in the quality of text classification and reducing the number of used indexes in comparison with previous works.