The rapid advancement of Large Language Models (LLMs) has revolutionized natural language processing (NLP), enabling precise automated topic generation and keyword extraction. This study proposes a framework integrating LLMs with topic modeling techniques to enhance thematic analysis in academic research. By employing clustering methodologies such as BERTopic, HDBSCAN, and KeyBERT, we evaluate the efficacy of automated topic modeling in structuring large-scale textual datasets. Analyzing over 5000 Web of Science-indexed publications (2017–-2023), we assess the impact of LLM-driven models on thematic coherence, reproducibility, and keyword precision compared to traditional methods. Empirical findings demonstrate that the proposed approach significantly enhances topic modeling efficiency across disciplines, including social sciences, market analysis, and industrial research. This research advances LLM-based NLP by offering a scalable, data-driven solution for systematic knowledge extraction, reinforcing the transformative role of LLMs in research synthesis and knowledge management.