Abstract: Background　Traditional two-dimensional imaging has limitations in the assessment of liver trauma, such as significant inter-observer variability and insufficient quantification of irregular trauma. The development of AI-assisted threedimensional imaging technology offers new possibilities for objective and quantitative assessment, yet its application value in the diagnosis and treatment of liver trauma has not been systematically verified. Objective　To incorporate AI-assisted 3D imaging technology into the liver trauma diagnosis and treatment workflow, and evaluate its feasibility and clinical value in terms of grading accuracy, correlation with trauma scores, accuracy in predicting final treatment outcomes, and efficiency of quantitative measurement. Methods　The clinical and imaging data of 109 liver trauma patients admitted from January 1, 2014, to September 1, 2025 in the First Medical Center of PLA General Hospital were retrospectively analyzed. All cases were graded by clinicians based on both 2D and 3D images according to the 2018 American Association for the Surgery of Trauma (AAST) liver trauma grading standards, and the differences in AAST grading results between the two methods were compared. Using the patient's treatment outcome as the gold standard, the receiver operating characteristic (ROC) curve was employed to evaluate the accuracy of both methods in predicting treatment outcomes, and predictive efficacy parameters were calculated using paired four-fold tables. Additionally, the consistency and time consumption between 3D reconstruction and 2D manual contouring in measuring liver volume, trauma volume, and hemoperitoneum volume were compared. Results　A total of 109 patients were enrolled, including 76 males (69.72%), 33 females (30.28%), with the mean age of (40.98±15.62) years. Traffic accidents (62.39%) were the main cause of trauma. Seventy-eight (72%) patients received conservative treatment, and 31 (28%) underwent surgery. The number of cases graded as AAST I-V by 2D imaging were 22, 25, 24, 28, and 10, respectively; while those assessed by 3D imaging were 22, 29, 28, 18, and 12, respectively. Regarding trauma severity classification, compared with 2D imaging, the number of severe trauma patients identified by 3D imaging decreased from 38 to 30, while mild-to-moderate trauma patients increased from 71 to 79. Significant differences were observed in abdominal ISS, AIS, and APACHE Ⅱ scores between the mild-to-moderate and severe trauma groups in both 3D and 2D imaging (P＜0.05). Regarding treatment outcomes, the proportion of patients receiving surgical treatment within the severe trauma group increased from 44.74% (based on 2D classification) to 63.33% (based on 3D classification). The area under the curve (AUC) for 3D AAST grading in predicting final treatment outcomes was 0.773, which was higher than the 0.707 for 2D grading. Volume measurement by AI-assisted 3D imaging showed high consistency with 2D manual contouring but required significantly less time (29.56± 11.46 min vs 57.16 ± 17.32 min, P＜0.001). Conclusion　AI-assisted 3D imaging technology can improve upon the limitations of traditional 2D imaging in liver trauma assessment, achieving a better match between grading and actual treatment strategies while significantly enhancing the efficiency and objectivity of quantitative assessment. As a feasible and innovative assessment tool, 3D imaging is expected to serve as a beneficial supplement to the liver trauma diagnosis and treatment workflow.