Objective To evaluate left univentricular (LUV) pacing for cardiac resynchronization therapy (CRT) using a rate-adaptive atrioventricular delay (RAAVD) algorithm to track physiological atrioventricular delay (AVD). Methods A total of 72 patients with congestive heart failure (CHF) were randomized to RAAVD LUV pacing versus standard biventricular (BiV) pacing in a 1: 1 ratio. Echocardiography was used to optimize AVD for both groups. The effects of sequential BiV pacing and LUV pacing with optimized A-V (right atrio-LV) delay using an RAAVD algorithm were compared. The standard deviation (SD) of the S/R ratio in lead V1 at five heart rate (HR) segments (RS/R-SD5), defined as the "tracking index," was used to evaluate the accuracy of the RAAVD algorithm for tracking physiological AVD. Results The QRS complex duration (132 +/- 9.8 vs. 138 +/- 10 ms, P < 0.05), the time required for optimization (21 +/- 5 vs. 50 +/- 8 min, P < 0.001), the mitral regurgitant area (1.9 +/- 1.1 vs. 2.5 +/- 1.3 cm(2), P < 0.05), the interventricular mechanical delay time (60.7 +/- 13.3 ms vs. 68.3 +/- 14.2 ms, P < 0.05), and the average annual cost (13,200 +/- 1000 vs. 21,600 +/- 2000 RMB, P < 0.001) in the RAAVD LUV pacing group were significantly less than those in the standard BiV pacing group. The aortic valve velocity-time integral in the RAAVD LUV pacing group was greater than that in the standard BiV pacing group (22.7 +/- 2.2 vs. 21.4 +/- 2.1 cm, P < 0.05). The RS/R-SD5 was 4.08 +/- 1.91 in the RAAVD LUV pacing group, and was significantly negatively correlated with improved left ventricular ejection fraction (LVEF) (Delta LVEF, Pearson's r = -0.427, P = 0.009), and positively correlated with New York Heart Association class (Spearman's r = 0.348, P = 0.037). Conclusions RAAVD LUV pacing is as effective as standard BiV pacing, can be more physiological than standard BiV pacing, and can decrease the average annual cost of CRT.