Objective: Addressing the compatibility issues between current Diagnosis-Related Group (DRG) payment rules and Traditional Chinese Medicine (TCM) clinical practice, this study focuses on the “dual-track” DRG practice in Chongqing. It aims to quantitatively map the core risk factors triggering DRG financial losses in different-tier TCM institutions through intelligent sensitivity testing, providing targeted empirical evidence for the differentiated optimization of medical insurance policies. Methods: A progressive research framework combining “macro-level empirical qualitative analysis and micro-level algorithmic simulation” was constructed. (1) Field research was conducted on 26 multi-tier TCM institutions in Chongqing via stratified random sampling to obtain baseline characteristics and execution pain points. (2) The Monte Carlo algorithm was introduced to generate a virtual micro-medical record dataset of 5,000 cases based on the surveyed macro-statistical distribution. (3) An intelligent sensitivity testing model based on the Random Forest algorithm was built with “whether a single case incurred a DRG settlement loss” as the dependent variable. This model quantitatively mapped the sensitivity (feature importance) of various clinical features causing losses under established rule constraints, with its efficacy verified by ROC curves and AUC values. Results: Chongqing’s “dual-track” reform had initially consolidated the positioning of grassroots TCM services, however, structural frictions persist at the micro-implementation level. The model’s sensitivity analysis revealed that tertiary hospitals were highly sensitive to the “TCM treatment rate” assessment, and their losses were primarily driven by the rigid reliance on modern medical interventions for severe cases (highest sensitivity contribution score). Conversely, secondary hospitals were highly sensitive to the “average length of stay,” where prolonged rehabilitation leading to fixed-quota cost inversion constitutes the core financial risk. Additionally, execution pain points such as the undervaluation of core TCM technologies were corroborated through field surveys. Conclusion: It is recommended to implement categorized optimization strategies based on the intelligent sensitivity test results: exploring flexible exemptions for the “TCM treatment rate” in specific intensive care departments and special-case negotiation mechanisms for tertiary hospitals; promoting a “DRG fixed quota + flexible bed-day” mixed payment model for secondary hospitals; and strengthening underlying data cross-system mapping and rational pricing mechanisms to further enhance the refinement and inclusivity of the TCM DRG payment system.