The current issues concerning the ancient tree resources along Beijing Central Axis include fragmented multi-source data, prominent semantic heterogeneity, insufficient exploration of cultural value, and weak representation of knowledge associations, which hinder the digital preservation and dissemination of ancient tree cultural heritage. To achieve the structured integration and intelligent application of multi-dimensional knowledge related to the nature, history, culture, and ecology of ancient trees, this study focuses on ancient trees along Beijing Central Axis. It integrates multi-source data acquisition methods such as laser scanning, satellite remote sensing, and OCR-based archive recognition. Based on natural language processing (NLP) techniques, entity extraction and relation annotation are carried out, and data cleaning and standardization are performed to address semantic heterogeneity. An extended ontology model comprising eight core concepts (individual tree, protection information, value dimensions, spatiotemporal data, people, institutions, and literature) is constructed. Knowledge storage, visual querying, and semantic reasoning are implemented using the Neo4j graph database. The results show that the constructed knowledge graph can clearly present the association paths of ‘ancient tree-tree species-people-events-locations-value’, demonstrating strong interactive visualization and intuitive cultural expression, effectively supporting precise retrieval of ancient tree resources and mining of implicit knowledge. The study suggests that further integration of multi-source data and optimization of reasoning mechanisms are key pathways to improving the coverage and accuracy of the knowledge graph. It also recommends establishing industry standards for data sharing and strengthening regional cooperation to provide a replicable paradigm for digital projects focused on ancient tree culture.