Objective To construct Logistic regression model and score system of early endometrial carcinoma based on two-dimensional ultrasound, and to discuss its clinical application value.Data and Methods 475 patients with endometrial lesions confirmed by surgery and pathology in our hospital were randomly divided into a training set (333 cases) and a verification set (142 cases) according to 7:3, and were divided into a benign group and an early malignant group according to pathology. Two-dimensional ultrasound examination was performed before surgery to compare the differences in two-dimensional ultrasound, serological tumor markers and clinical indicators between the two groups.Binary Logistic regression analysis was used to screen the independent influencing factors of early malignant transformation in patients with endometrial lesions, and the Logistic regression model and scoring system were constructed for further internal verification.The receiver operating characteristic (ROC) curve was drawn to analyze the model and the distinguishing ability of the scoring system. The calibration ability of the model was evaluated by Hosmer-Lemeshaw (H-L) test. The U-test was used for ROC curve analysis and comparison.Results Binary Logistic regression analysis showed that vaginal bleeding, menopause, endometrial thickening, uterine effusion, unclear boundary between endometrium and muscle, and rich blood supply were independent risk factors for early endometrial cancer (P < 0.05).The model and scoring system were constructed based on the above six indexes. The ROC curve analysis showed that the AUC was 0.935 (95%CI0.903-0.966), indicating good calibration ability (P=0.868>0.05).ROC curve analysis of the scoring system showed that the AUC was 0.914 (95%CI0.874-0.954).The diagnostic sensitivity and specificity were 79.76% and 91.57%, respectively, with a total score of 2.5.The differential diagnosis efficiency of the model and the scoring system was similar in the training set and the verification set (P = 0.256, P = 0.481).The scoring system could replace the model to predict the risk of early EC in patients with endometrial disease (P=0.062).Conclusion Logistic regression model and scoring system based on two-dimensional ultrasound have good clinical value in predicting early malignant transformation in patients with endometrial disease.