For patients with atrial fibrillation (AF) who are refractory to anti-arrhythmic drugs (AADs), minimally invasive video-assisted thoracoscopic surgical ablation (SA) and catheter ablation (CA) are potential alternative treatment options. The recent FAST randomized study suggested that thoracoscopic SA was superior to CA in achieving freedom of AF in patients who have failed at least one prior AAD. To assess the relative merits and risks of SA versus CA, a systematic review and meta-analysis was conducted. Electronic searches were performed using six databases from their inception to December 2014. Relevant studies comparing thoracoscopic SA and CA were identified; data were extracted and analysed according to predefined clinical endpoints. Relative risk (RR) and weighted mean difference were used as summary statistics. Freedom from AF/arrhythmias was significantly higher in SA versus CA at 12-month off-AAD (78.4 vs 53%; RR, 1.54; P < 0.0001) and on-AAD (82.6 vs 45.7%; RR, 1.85; P < 0.00001). This difference was maintained in paroxysmal and persistent AF subgroups. The SA cohort had a significantly lower requirement for repeat ablations compared with the CA cohort (4.7 vs 24.4%; RR, 0.21; P = 0.0001). However, major complications were significantly higher in the SA group (28.2 vs 7.8%; RR, 3.30; P = 0.0003), driven by pleural effusion and pneumothorax. SA may be more efficacious than CA treatment in a selected patient population with refractory AF and prior failed catheter intervention. Improved freedom from arrhythmias at up to 12-month follow-up is counterbalanced by higher procedural complication rates.