Over the years, geological surveys and information infrastructure development have accumulated vast amounts of multi- source, heterogeneous geoscience data. However, traditional geological information processing relies on manual data entry and fragmented analysis, resulting in low data integration efficiency and delayed decision- making. To address the challenge of low utilization efficiency for unstructured geological texts, this study designed an intelligent information extraction and visualization system.Methods: The system adopts a frontend- backend separation architecture. Multiple large language models are introduced as the core information extraction backbone. Domain knowledge is incorporated through user- defined templates to guide the extraction of structured information from multimodal geological data. Extracted geological entities and attributes are represented using representation learning techniques. Similarity measurement methods are applied to establish multi- feature associations among geological data, based on which a geological knowledge graph is constructed. Multidimensional interactive visualization is implemented using ECharts.Results: This integrated platform achieves transformation from text to knowledge, automatically generating knowledge graphs to reveal entity relationships. It provides an effective solution for intelligent processing and deep utilization of geological texts. By integrating cutting- edge artificial intelligence and visualization technologies, it significantly enhances the usability and discovery efficiency of geological knowledge, laying a solid practical foundation for building an intelligent geological knowledge service platform.Conclusions: This study focuses on intelligent extraction and visualization of unstructured geological texts and presents the design and implementation of an information extraction and visualization system for geological applications. Overall, the proposed system provides a feasible implementation pathway for the deep mining and utilization of geological textual data.