<p>This study offers a comprehensive mapping of geoscience research from 2000 to 2024, using an integrated bibliometric and topic modelling methodology. Using a carefully chosen set of 4513 Scopus-indexed journal articles, Latent Dirichlet Allocation identified 15 coherent research themes. These encompass fundamental fields such as mineral resource exploration, tectonic structure analysis, and climate change, as well as emerging domains including machine learning, spatial data analysis, digital information systems, and geoscience education. The results show that scientific productivity has steadily increased, particularly since 2010, and that the global collaboration network has expanded. The United States, the United Kingdom, Germany, and China were among the most important contributors. Belgium and Norway, on the other hand, had high citation efficiency despite publishing fewer articles. Patterns of topic co-occurrence and keyword clustering indicate that computational tools and traditional geoscientific methods are converging. This comprehensive analysis underscores the dynamic and interdisciplinary nature of contemporary geoscience, offering fresh perspectives on research deficiencies, shifting priorities, and the discipline’s structural evolution. The findings facilitate informed decision-making for researchers, institutions, and policymakers, underscoring the importance of technology-driven, collaborative strategies to address complex environmental and geological challenges.</p>