The research focuses on the automation of intellectual processing of Internet content for the timely detection and blocking of content operating destructively on the user's mind. It deals with subjects of current interest (such as pressure upon mentality of the user, incigation, cyberbullying, incitement of the masses, etc.) and connects several scientific directions. This article shows results of the research in the field of text-based Internet content analysis. Numerical indicators of the estimated emotional components for content with signs of aggression and mental pressure upon the user are shown in work. The obtained results of the lexical analysis of electronic texts demonstrate the presence of a natural increase in the proportion of use of negatively coloured lexemes in the texts with signs of aggression. Also, the structure of the socio-cyberphysical system of analyzing Internet content with a description of key blocks and elements is presented.