This article explores the application of Natural Language Processing (NLP) techniques and the Fuzzy c-means clustering algorithm for the automated categorization of social media posts based on expert-defined topics. The study analyzes data from Avito, VKontakte communities, including Volga-IT, UlSTU Students Community, Ulyanovsk State Technical University, and RAII Summer School. The initial stage performs text preprocessing to extract keywords, terms, and named entities, then employs a large language model to identify thematic categories. The Fuzzy c-means algorithm is then applied to cluster posts, accounting for their partial membership in multiple topics, which reflects the inherent ambiguity of social media content. The results of automated clustering are compared with expert-labeled data to evaluate accuracy and consistency. The proposed approach can benefit SMM specialists by providing a tool for automated content analysis, reducing manual categorization efforts, and improving community management efficiency. Key advantages of this method include its flexibility in handling vaguely defined topics and compatibility with modern NLP technologies. The study demonstrates the potential of combining fuzzy logic and machine learning for enhancing social media analytics, offering a practical solution for digital content classification.