Background: Multisite pacing has been suggested to be effective in suppression of atrial fibrillation (AF) by shortening of atrial activation time and multidirectional excitation. However, the underlying electrophysiologic mechanisms remains unclear, and the efficacy of multisite pacing in human has not been evaluated. The purpose of this study is to compare the effect of linear multi- to single-site pacing on activation time (AT) and prevention of AF. Methods: In 7 patients (Persistent AF:Paroxysmal AF=5:2, 5 male, age 56.3짹13.5 years old, left atrium (LA) size 51.1짹12.6 mm, left ventricular ejection fraction 56.4짹6.3 %), noncontact endocardial mapping was performed using a multielectrode array positioned at the (LA). Multi- and single-site pacing were performed at the high righ atrium (HRA), distal coronary sinus (CS), and the LA during AF and sinus rhythm. The pacing site at the LA was determined by multiple overlay on dynamic substrate map (DSM). AF inducibility (50 ms, 5 sec, 10 mA) was compared during continuous single vs. triple sites of pacing (500-600 ms) at each site. Atrial AT during sinus rhythm and propagation pattern during AF were also compared. Results: Left- and bi-atrial AT was further shortened by multisite- than single-site pacing, and further decreased during distal CS pacing (LA AT: 86.4짹18.3 ms vs. 74.7짹12.0 ms , p=0.013; bi-atrial AT: 165.6짹29.4 ms vs. 144.1짹42.8 ms, p=0.029; right atrial AT: 106.7짹32.7 ms vs. 107.7짹38.3 ms, p=0.765). In 6 patients, wave break-up during AF mostly occurred at the multiple overlay sites on DSM, and in 4 patient, complex fractionated atrial electrograms (CFAEs) were recorded at the same site of wave break-up. Triple-site pacing at the LA where CFAEs and wave break-up were noted, terminated AF in 2 patients and prevented AF induction in 1 patient. In 2 patients, AF was organized by LA pacing (sigle 1, triple 1). Conclusion: Linear multisite pacing, specifically at the LA showing wave break-up or CFAEs, suppressed AF by multidirectional excitation and minimization of bi-atrial activation time. Further studies including large number of the patients are warranted to clarify this finding.