Establishing referential relations in discourse is one of the most relevant but difficult to model problems of automatic text analysis. Reference means attributing a textual unit (linguistic expression) to a non-linguistic object (referent). The correct interpretation of the statement in the analyzed text requires solving the referent of the textual reference about the object, that is, the reference of the textual expression. In this work, a model for resolving referential relations is proposed. The method used to solve the referential relationship is finding equivalence, that is, by calculating the degree of similarity of possible elements in the word, the similarity in the referential relationship is calculated, and as a result, the relationship value of the referent in the sequence of sentences is obtained. These opportunities will be a new beginning for solving referential relationships in the Kazakh language. Based on the possibilities of computer technology, this direction in developing language education creates a great opportunity for extensive research and analysis at various levels of the language treasure as a bright mirror of the spiritual and material culture of the people.