This thesis explores the representation of polysemy, synonymy, and homonymy in the Uzbek electronic thesaurus, emphasizing their role in structuring lexical resources for natural language processing (NLP) and information retrieval. The study examines methods for distinguishing word meanings, organizing synsets, and integrating computational approaches such as AI and word embedding models. It highlights the importance of linguistic validation and hierarchical organization in thesaurus development. By improving semantic representation, the Uzbek electronic thesaurus enhances machine translation, search systems, and AI-driven text processing applications.