Suicide is one of the top external factors contributing to deaths particularly in youths. Therefore it arises a need to come up with methods to identify suicidal tendencies in individuals. Speech analysis, particularly quantitative one, is of great significance. Scientists are currently developing mathematical models to enable identification of suicidal tendencies of individuals based on numerical values of linguistic parameters of their written texts. This research is conducted using English materials. The article describes methods and approaches employed in the identification of suicidal tendencies in authors of written texts in Russian. This is the first time this problem has been tackled. It is not just the research material, which is a specially designed corpus of texts by suicidal individuals, that gives the problem a new dimension, but it is also the approach to choosing text parameters to analyze based on self-destructive theories currently dominating psychology, neurobiology of suicidal behaviour on one hand and neurolinguistic data on cerebral mechanisms of meaningful speech production as well as in terms of brain asymmetry and formulation of a mathematical task of this particular research on the other. It is shown that using the suggested methods and approaches new scientific data on language predictors of suicidal behaviour typical of individuals in Russian language (at different levels) can be obtained as well changes of correlation text parameters can be identified as suicidal tendencies progress.