Objective: To compare the effects of biventricular electrical pacing and conventional single-ventricular pacing for cardiac contractility modulation (CCM) on cardiac contractile function and to delineate the underlying molecular mechanisms. Methods: Forty rabbits were divided into four groups before surgery: healthy control, HF sham, HF left ventricular pacing CCM (LVP-CCM), and HF biventricular pacing CCM (BVP-CCM) groups with n=10 for each group. A rabbit model of chronic heart failure was established by ligating ascending aortic root of rabbits. Then electrical stimulations during the absolute refractory period were delivered to the anterior wall of left ventricle in the LVP-CCM group and on the anterior wall of both left and right ventricles in the BVP-CCM group lasting six hours per day for seven days. Changes in ventricular structure, cardiac function and electrocardiogram were monitored before and after CCM stimulation. Results: Compared with the sham-operated group, heart weight, heart weight index, LV end-systolic diameter (LVESD), LV end-diastolic diameter (LVEDD) in the LVP-CCM and BVP-CCM groups were significantly decreased (p<0.05), while LV ejection fraction (LVEF) and fractional shortening fraction (FS) were increased (p<0.05). Notably all these changes were consistently found to be greater in BVP-CCM than in LVP-CCM. Moreover, plasma BNP levels were highest in the HF sham-control group, followed by the LVP-CCM group, and lowest in the BVP-CCM group (p<0.05). Furthermore, sarco/endoplasmic reticulum Ca2+-ATPase (SERCA2a) protein levels were upregulated by 1.7 and 2.4 fold, along with simultaneous upregulation of a cardiac-enriched microRNA miR-133 levels by 2.6 and 3.3 fold, in LVP-CCM and BVP-CCM, respectively, compared to sham. Conclusions: Biventricular pacing CCM is superior to conventional monoventricular pacing CCM, producing greater improvement cardiac contractile function. Greater upregulation of SERCA2 and miR-133 may account, at least partially, for the improvement by BVP-CCM.