This paper explores the application of AI technology in cultural heritage data management, focusing on wall paintings’ condition assessment data from Mogao Grottoes, and constructs a framework integrating graph data structures with an AI model. The method integrates multi-source data from Mogao Grottoes wall paintings surveys, such as handwritten records and digital archives, to facilitate efficient analysis and rapid query of deterioration and spatio-temporal information. Leveraging this novel technical framework, the study enhances the intelligence of cultural heritage data management, offering valuable approaches for the conservation of similar heritage sites. The findings effectively advance the digital and intelligent transformation of cultural heritage conservation, aligning with the focus on data-driven diagnosis for conservation decision-making.