Background: Low radiofrequency powers are commonly used on the posterior wall of the left atrium for atrial fibrillation ablation to prevent esophageal damage. Compared with higher powers, they require longer ablation durations to achieve a target lesion size index (LSI). Esophageal heating during ablation is the result of a time-dependent process of conductive heating produced by nearby radiofrequency delivery. This randomized study was conducted to compare risk of esophageal heating and acute procedure success of different LSI-guided ablation protocols combining higher or lower radiofrequency power and different target LSI values. Methods: Eighty consecutive patients were prospectively enrolled and randomized to one of 4 combinations of radiofrequency power and target LSI for ablation on the left atrium posterior wall (20 W/LSI 4, 20 W/LSI 5, 40 W/LSI 4, and 40 W/LSI 5). The primary end point of the study was the occurrence and number of esophageal temperature alerts per patient during ablation. Acute indicators of procedure success were considered as secondary end points. Long-term follow-up data were also collected for all patients. Results: Esophageal temperature alerts occurred in a similar proportion of patients in all groups. Significantly, shorter radiofrequency durations were required to achieve the target LSI in the 40 W groups. Less than 50% of the radiofrequency lesions reached the target LSI of 5 when using 20 W despite a longer radiofrequency duration. A lower rate of first-pass pulmonary vein isolation and a higher rate of acute pulmonary vein reconnection were recorded in the group 20 W/LSI 5. A lower atrial fibrillation recurrence rate was observed in the 40 W groups compared with the 20 W groups at 29 months follow-up. Conclusions: When guided by LSI, posterior wall ablation with 40 W is associated with a similar rate of esophageal temperature alerts and a lower atrial fibrillation recurrence rate at follow-up if compared with 20 W. These data will provide a basis to plan future randomized trials. Registration: URL: https://www.clinicaltrials.gov . Unique identifier: NCT02619396.