Introduction: We analyzed the influence of the QRS duration (QRSd) to left ventricle (LV) end-diastolic volume (LVEDV) ratio on the clinical outcome of cardiac resynchronization therapy (CRT) in HF patients with small body size and New York Heart Association (NYHA) classification III/IV. Methods and Results We enrolled 114 patients with advanced heart failure (NYHA class III/IV, and LV ejection fraction > 35%) who received a CRT device, including those with left bundle branch block (LBBB) and QRSd ≥ 120 ms (n = 60), non-LBBB and QRSd ≥ 150 ms (n = 30), and non-LBBB and QRSd of 120-149 ms (n = 24). Over a mean follow-up period of 65 ± 58 months, the incidence of the primary endpoint, a composite of all-cause death and hospitalization for heart failure, showed no significant intergroup difference (43.3% vs. 50.0% vs. 37.5%, respectively, p = 0.72). Similarly, among 104 patients with QRSd/LVEDV ≥ 0.67 (n = 54) and QRSd/LVEDV < 0.67 (n = 52), no significant differences were observed in the incidence of the primary endpoint (35.1% vs. 51.9%, p = 0.49). Nevertheless, patients with QRSd/LVEDV ≥ 0.67 showed better survival than those with QRSd/LVEDV < 0.67 (14.8% vs. 34.6%, p = 0.0024). Conclusion: Advanced HF patients with a higher QRSd/LVEDV ratio showed better survival in this small-body–size population. Thus, the risk is concentrated among those with a larger QRSd, and patients with a relatively smaller left ventricular size appeared to benefit from CRT.