Background: Myocardial strain rate and torsion are useful tools to measure left ventricular (LV) contractility. However, clinical values of these parameters in predicting LV systolic function after surgical correction of severe mitral regurgitation (MR) have not been demonstrated. Methods: In 68 patients with severe MR, 2D echo and speckle tracking imaging were performed at ��� 7 days before surgery, and longitudinal, radial, and circumferential strains and peak strain rates, and LV torsion were measured. After surgery, echo was repeated within 7 days. Results: Post-surgery LV ejection fraction (EF) significantly correlated with pre-surgery LV end-systolic (r=-0.65) and end-diastolic (r=-0.35) dimensions, end-systolic (r=-0.55) and end-diastolic (r=-0.55) volumes, EF (r=0.36), MR orifice area (r=-0.39), and circumferential (r=-0.41) and longitudinal (r=-0.26) strain rates, whereas it did not with strains and LV torsion. By multiple linear regression, pre-surgery LV end-systolic dimension (p<0.001), circumferential strain rate (p<0.05), and MR orifice area (p<0.01) were independent predictors of post-surgery EF. ROC curve for predicting a normal post-surgery LV EF (��� 50%) showed that area under curve of end-systolic LV dimension (0.81) was larger than those of circumferential strain rate (0.71) and MR orifice area (0.70). Pre-surgery end-systolic LV dimension ��� 42 mm predicted a normal post-surgery LV EF with a sensitivity of 88% and a specificity of 59%. Conclusion: LV circumferential peak strain rate was an independent predictor, but speckle tracking parameters were not superior to conventional parameters in predicting LV EF after surgery of severe MR.