The association between objectively measured physical activity and atrial fibrillation (AF) has not been examined. Therefore, we examined the association between moderate and vigorous physical activity (MVPA) with incident AF in 5,147 participants who completed accelerometer assessment for 4 to 7 consecutive days in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study. MVPA was defined as >1,065 counts/minute, and daily mean time spent in MVPA was computed. Incident AF was identified during follow-up by a study-scheduled electrocardiogram and also from self-reported medical history of a physician diagnosis. Logistic regression was used to assess the relation between daily time in MVPA and incident AF. A total of 429 (8.3%) incident AF cases were detected after a median follow-up of 3.5 years following accelerometer assessment. Participants with higher daily time spent in MVPA were less likely to develop AF than those with lower MVPA achievement (Quartile 1 = 12.4%; Quartile 2 = 8.3%; Quartile 3 = 7.1%; Quartile 4 = 5.4%; p-trend <0.001). In a multivariable model adjusted for AF risk factors, the risk of AF decreased with higher levels of daily MVPA (Quartile 1: Ref; Quartile 2: hazard ratio [HR] = 0.77, 95% confidence interval [CI] 0.58 to 1.01; Quartile 3: HR = 0.72, 95% CI 0.53 to 0.98; Quartile 4: HR = 0.62, 95% CI 0.44 to 0.87; p-trend = 0.0056). In conclusion, higher levels of objectively measured daily MVPA are protective against the development of AF suggesting that promotion of MVPA should be encouraged to reduce the risk of AF.