Survival difference following lobectomy, segmentectomy, and wedge resection in stage I non-small cell lung cancer (NSCLC) and its subgroups remains undetermined. We systemically searched published articles comparing recurrence-free survival (RFS), overall survival (OS), or cancer-specific survival (CSS) between lobectomy and limited resection or between segmentectomy and wedge resection. A total of 42 studies published from 1980 to 2014 enrolling 21,926 patients were included in this meta-analysis. Survival results favored lobectomy in stage IA NSCLC ≤2 cm (combined HR: 1.530, 95% CI: 1.402-1.671, P < 0.001) or patient's ≥65 years old (combined HR: 1.227, 95% CI: 1.003-1.502, P = 0.047). Survival outcome of video-assisted thoracoscopic (VATS) sublobectomy was comparable to that of VATS lobectomy (pooled HR: 0.808, 95% CI: 0.556-1.174, P = 0.263). The combined HR of segmentectomy versus lobectomy was 1.231 (95% CI: 1.070-1.417, P = 0.004), while the pooled HR of wedge resection versus segmentectomy was 1.542 (95% CI: 0.856-2.780, P = 0.149). This study suggested that tumor size or age alone should not be the criteria to encourage sublobar resection. For stage I NSCLC, survival following segmentectomy was inferior to lobectomy. Patients undergoing intentional sublobectomy achieved comparable survival as those who received lobectomy.