This study was designed to validate new parameters based on the relationship between RR interval and left ventricular(LV) performance, representing LV function independent of irregular RR intervals in atrial fibrillation(AF). Echocardiography was performed in 107 consecutive patients with AF. The logarithmic equation between pre-preceding RR interval(RR-2) and LV outflow peak ejection velocity(Vpe) was obtained in the coordinates with preceding RR interval(RR-1)���0.5 second. By removing the coordinates RR-1< 0.5 second, the squared correlation coefficient(r²) between RR-2 and Vpe was improved from 0.19 to 0.25 (p=0.000). Using this equation, Vpe was adjusted for the effect of RR-2 assuming that RR-2 is fixed to the mean RR interval. By this adjustment, r² between RR-1 and Vpe was improved from 0.61 to 0.71 (p=0.000). The logarithmic equation between RR-1 and adjusted Vpe was calculated in the coordinates with RR-1 ranged from 0.6 to 1 second. From this relationship, r², slope, Vpe at RR-1 = 1 second(Vpe-1), and the ratio of slope to Vpe-1(Slope/Vpe-1) were calculated. When patients were divided into two groups according to the history of heart failure(HF, NYHA FC ���3, n=54), forward logistic regression analysis showed that old age, high Slope/Vpe-1, mitral regurgitation, and LA enlargement independently predict the occurrence of HF. Ejection fraction and fractional shortening were not different between two groups. Slope/Vpe-1 was independently associated with systolic LV dimension and history of heart failure (r=0.49, p=0.000). To minimize the effect of systolic dysfunction, same analyses were repeated in patients with normal LV size and without significant regurgitation (n=69). Old age and high Slope/Vpe-1 independently predicted the occurrence of HF (n=27). Slope/Vpe-1 was independently associated with thickness of interventricular septum and the history of HF (r=0.45, p=0.001). In summary, a new parameter, Slope/Vpe-1, is superior to the classic parameters for the evaluation of LV function. This parameter is determined not only by the systolic dysfunction but also by the diastolic dysfunction resulting in the most useful predictor in predicting the occurrence of HF in AF.