Abstract: Current drilling operations face challenges such as dispersed multi-source data, severe information silos, and the disconnection between design and field implementation, which lead to insufficient decision-making support, low operational efficiency, and delayed risk response. Digital twin technology provides a feasible approach to addressing these issues. However, existing wellbore digital twin models still suffer from limitations in real-time data updating, multi-source data integration, and business process coupling, making it difficult to meet the requirements of precise and real-time intelligent decision-making. To address these problems, a knowledge-embedded digital twin modeling method for drilling wellbores is proposed. Centered on a wellbore stability mechanics model, the method integrates mechanics-based modeling, data-driven algorithms, and natural language processing techniques, and incorporates ontology, knowledge graphs, and rule bases to achieve semantic enhancement and alignment of multi-source information. A multi-layer architecture including data acquisition, storage, modeling analysis, and decision support is constructed. The proposed approach enables real-time mapping of wellbore states and automatically parses unstructured records such as drilling design documents to generate structured knowledge and optimize operational workflows. A case study on a coalbed methane well is conducted to validate the method in wellbore stability prediction, ROP optimization, and automated operation instruction generation. The results show that the model can monitor downhole dynamics in real time, accurately predict potential risks, and generate optimized decision-making schemes. The ROP in the test interval increased by 12.1%, and the efficiency of operational document processing improved by more than 80%. The proposed approach significantly enhances the perception, prediction, and operational applicability of digital twin systems, providing a new technical pathway for safety assurance, efficiency improvement, and intelligent decision-making in complex drilling operations.