We tested the hypothesis that left ventricular (LV) pacing is superior to right ventricular (RV) apical pacing in patients undergoing atrioventricular (AV) junction ablation and pacing for permanent atrial fibrillation. The potential benefit of LV over RV pacing needs to be evaluated without the confounding effect of other variables that can influence cardiac performance. An acute intrapatient comparison of the QRS width and echocardiographic parameters between RV versus LV pacing was performed within 24 h after ablation in 44 patients. Both modes of pacing were also compared with pre-implantation values. Compared with RV pacing, LV pacing caused a 5.7% increase in the ejection fraction (EF) and a 16.7% decrease in the mitral regurgitation (MR) score; the QRS width was 4.8% shorter with LV pacing. Similar results were observed in patients with or without systolic dysfunction and/or native left bundle branch block, except for a greater improvement in MR in the latter group. Compared with pre-ablation measures, the EF increased by 11.2% and 17.6% with RV and LV pacing, respectively; the MR score decreased by 0% and 16.7%; and the diastolic filling time increased by 12.7% and 15.6%.Rhythm regularization achieved with AV junction ablation improved EF with both RV and LV pacing; LV pacing provided an additional modest but favorable hemodynamic effect, as reflected by a further increase of EF and reduction of MR. The effect seems to be equal in patients with both depressed and preserved systolic functions and in those with and without native left bundle branch block.