This article is focused on the problem of translating literary discourse texts using machine translation. The analysis of translations of literary texts has shown certain difficulties for machine translation in defining traditional problematic phenomena such as homonymy and synonymy, nationally marked and emotionally expressive units, derivational formants of intensification, subjunctive mood predicative constructions, elliptical constructions. The purpose of the study is to analyze the use of automated translation programs in teaching the translation of literary texts; to formulate methodological recommendations aimed at compensating translation difficulties of automated translation programs, to develop methodological recommendations on the use of a combined method of translating literary texts.