Surgical ablation of atrial fibrillation (AF) has become a routine procedure during concomitant cardiac surgery, however, the extension of lesion sets remain controversial. We sought to compare the relative benefit and risk of different lesion sets through a Bayesian network meta-analysis (NMA).Pubmed, Embase, and Cochrane Trials databases were searched for randomized controlled trials (RCTs) comparing the rhythm outcome of AF patients undergoing pulmonary vein isolation (PVI), left atrial Maze (LAM), bi-atrial Maze (BAM), or no ablation during concomitant cardiac surgery. An NMA was conducted to explore the difference of over 1 year AF freedom as well as risks for early mortality and permanent pacemaker implantation (PPMI).A total of 2031 patients of 19 RCTs were included. PVI, LAM, and BAM (OR [95% Cr.I]: 5.02 [2.72, 10.02], 7.97 [4.93, 14.29], 8.29 [4.90, 14.86], p < .05) demonstrated higher freedom of AF compared with no ablation, however, no significant difference of rhythm outcome was found among the three ablation strategies based on the random-effects model. BAM was associated with an increase in early mortality when compared with no ablation (OR [95% Cr.I]: 4.08 [1.23, 17.30], p < .05), while none of the remaining comparisons reached statistical difference in terms of early mortality and PPMI.Bi-atrial ablation is not superior to left atrial ablation strategies in reducing AF recurrence for un-selected surgical patients. BAM has a higher risk of early mortality than no ablation, but no difference was found between bi-atrial and left atrial ablation in regard to early mortality and PPMI based on the current evidence.© 2021 Wiley Periodicals LLC.