Abstract Background The outcome of mitral valve repair over replacement in degenerative diseases has been recognized, but the patient selection of mitral valve repair in rheumatic diseases are still controversial. At present, there is a lack of comprehensive evaluation methods for rheumatic mitral valve disease, which cannot guide the selection of patients before surgery. Purpose The aim of this study was to establish a comprehensive evaluation method and innovative indicators for rheumatic mitral valve disease using transthoracic echocardiography, and to establish a model for predicting the outcome of rheumatic mitral valve repair. Methods This prospective, multicenter, observational cohort study included 167 patients who underwent rheumatic mitral valve surgery in 4 centers from 2022 to 2023. Each enrolled patient underwent a comprehensive transthoracic echocardiography evaluation and measured 94 ultrasound indicators, mainly covering leaflets, annulus, subvalvular apparatus, and atrial and ventricular function. The main endpoint event was the failure of mitral valve repair (including switching to replacement after repair, and moderate to severe stenosis or regurgitation during follow-up after repair). Patients were randomly assigned to the modeling and validation groups at a ratio of 7:3. In the modeling group, LASSO regression was used for variable selection, and logistic regression was used for modeling. A nomogram was drawn and a webpage rating calculator was created. In the validation group, model comparisons were conducted using the C-statistic and Net Reclassification Index (NRI). Results 112 patients (67.1%) underwent mitral valve repair with satisfactory results, 47 patients (28.1%) ultimately underwent mitral valve replacement, and 8 patients (4.8%) found moderate to severe mitral stenosis or regurgitation during follow-up after repair. Three indicators, including mitral valve orifice area, leaflet calcification score, and anterior leaflet-annulus angle, were selected through LASSO regression. No collinearity was found between the three indicators, and they were ultimately included in the logistic regression model. The C statistics in the modeling group and validation group were 0.846 and 0.958, respectively, indicating that the constructed model has satisfactory predictive ability. Compared to the Wilkins score, the proportion of correctly classified items increased by 19.8%, significantly better than the Wilkins score. Finally, a nomogram was drawn based on the model, and a web rating calculator was created to guide clinical decision-making. Conclusion This study explored a comprehensive evaluation method and innovative indicators for rheumatic mitral valve disease by transthoracic echocardiography, established and validated a model for predicting the outcome of repair surgery, and had certain guiding significance for doctors to choose suitable patients for rheumatic mitral valve repair.Validation of the predictive modelNomogram