In the article are reviewed research methods of the semantic structure in the spontaneous oral text. The semantic structure of the text is defined. Emphasis is placed on the choice of research method for the semantic structure in the text. Since the semantic structure of the text is the cognitive projection of the text in the human mind, therefore, its study must be based on psycholinguistic methods that allow use of the linguistic consciousness. However, when a large number of texts are investigated, for it is required large number of informants and it is very labor-intensive. Therefore, to continue a productive research of the semantic structure in the text, it is necessary to find a method that allows modeling of the semantic structure of texts without involvement of informants and adequately to their linguistic consciousness. That is why, are considered psycholinguistic method for research of the text integrity (method of topic allocation), method of graphosemantic modeling (basic method) and the method developed by the author of graphosemantic modeling considering syntactic relations (modification of basic methods). Analysis results acquired by processing the same text with these methods are compared. It is shown that the modified technique correlates with data acquired by psycholinguistic methods.