There is continuous debate regarding the effectiveness of thymectomy in the treatment of non-thymomatous myasthenia gravis (MG). This systematic review was undertaken to determine whether thymectomy was effective in non-thymomatous MG. We retrieved articles published between January 1980 and September 2013. Sixteen cohort studies were included. Given the considerable heterogeneity, we used a descriptive method instead of statistical synthesis. The median relative rates (RRs) and their interquartile ranges were used to estimate the magnitude of benefit. Compared to conservatively treated MG patients, thymectomized patients had higher survival, clinical remission, pharmacologic remission and improvement rates, and RRs were 1.07 (1.01-1.17), 1.83 (0.82-2.99), 1.55 (1.22-1.95) and 1 (1.00-1.09), respectively. Subgroup analyses showed that patients with moderate to severe generalized MG benefited more from thymectomy, with RRs of survival and pharmacologic remission increasing to 1.35 (1.24-1.49) and 2.68 (1.73-4.17), respectively. These results suggested that thymectomy might be an effective procedure in non-thymomatous MG patients. The patients with moderate to severe generalized MG might benefit more. Taking into account the poor methodological quality of present studies, more well-designed prospective randomized controlled trials (RCTs) are still required to reach unequivocal conclusion.