Aims The purpose of our study was to evaluate the feasibility and efficacy of cardiac resynchronization therapy (CRT) via left bundle branch pacing (LBBP-CRT) compared with optimized biventricular pacing (BVP) with adaptive algorithm (BVP-aCRT) in heart failure with reduced left ventricular ejection fraction <= 35% (HFrEF) and left bundle branch block (LBBB). Methods and results One hundred patients with HFrEF and LBBB undergoing CRT were prospectively enrolled in a non-randomized fashion and divided into two groups (LBBP-CRT, n = 49; BVP-aCRT, n = 51) in four centres. Implant characteristics and echocardiographic parameters were accessed at baseline and during 6-month and 1-year follow-up. The success rate for LBBP-CRT and BVP-aCRT was 98.00% and 91.07%. Fused LBBP had the greatest reduced QRS duration compared to BVP-aCRT (126.54 +/- 11.67 vs. 102.61 +/- 9.66 ms, P < 0.001). Higher absolute left ventricular ejection fraction (LVEF) and oLVEF was also achieved in LBBP-CRT than BVP-aCRT at 6-month (47.58 +/- 12.02% vs. 41.24 +/- 10.56%, P = 0.008; 18.52 +/- 13.19% vs. 12.89 +/- 9.73%, P = 0.020) and 1-year follow-up (49.10 +/- 10.43% vs. 43.62 +/- 11.33%, P = 0.021; 20.90 +/- 11.80% vs. 15.20 +/- 9.98%, P = 0.015, P = 0.015). There was no significant difference in response rate between two groups while higher super-response rate was observed in LBBP-CRT as compared to BVP-aCRT at 6 months (53.06% vs. 36.59%, P = 0.016) and 12 months (61.22% vs. 39.22%, P = 0.028) during follow-up. The pacing threshold was lower in LBBP-CRT at implant and during 1-year follow-up (both P < 0.001). Procedure-related complications and adverse clinical outcomes including heart failure hospitalization and mortality were not significantly different in two groups. Conclusions The feasibility and efficacy of LBBP-CRT demonstrated better electromechanical resynchronization and higher clinical and echocardiographic response, especially higher super-response than BVP-aCRT in HFrEF with LBBB.