The application of domestic turbodrills in drilling operations—particularly for enhancing drilling efficiency and enabling managed pressure drilling—has steadily increased in recent years. However, the failure modes associated with turbodrills are inherently complex and uncertain, posing significant operational risks when failures occur. Currently, turbodrill failure risk assessments in China predominantly rely on Fault Tree Analysis (FTA), a method that presents several limitations, such as the requirement for highly accurate failure probabilities, high computational demands, and potential delays in operational decision-making. To overcome these limitations, this study proposes a novel failure risk assessment approach based on Bayesian network theory, implemented using the GeNIe Academic software. Using the LD-101 well turbodrill as a case study, a Bayesian network model was developed, resulting in an initial calculated failure probability of 17%. A subsequent sensitivity analysis identified key contributing factors, and targeted mitigation strategies were applied. Following these interventions, the recalculated failure probability was reduced from 17% (moderate-to-high risk) to 9% (low risk).These findings confirm the feasibility and effectiveness of the proposed method and demonstrate its potential as a practical and data-driven tool for improving turbodrill risk assessment and management in drilling operations.