Background This study aimed to describe the incidence of atrial fibrillation (AF) among older adults during 3 years of follow-up. Methods and Results In this cohort study, 5201 adults ≥65 years old were examined annually on four occasions between June 1989 and May 1993. At baseline, participants answered questionnaires and underwent a detailed examination that included carotid ultrasound, pulmonary function tests, ECG, and echocardiography. Subjects with a pacemaker or AF at baseline (n=357) were excluded. New cases of AF were identified from three sources: (1) annual self-reports, (2) annual ECGs, and (3) hospital discharge diagnoses. Cox proportional-hazards models were used to assess baseline risk factors as predictors of incident AF. Among 4844 participants, 304 developed a first episode of AF during an average follow-up of 3.28 years, for an incidence of 19.2 per 1000 person-years. The onset was strongly associated with age, male sex, and the presence of clinical cardiovascular disease. For men 65 to 74 and 75 to 84 years old, the incidences were 17.6 and 42.7, respectively, and for women, 10.1 and 21.6 events per 1000 person-years. In stepwise models, the use of diuretics, a history of valvular heart disease, coronary disease, advancing age, higher levels of systolic blood pressure, height, glucose, and left atrial size were all associated with an increased risk of AF. The use of β-blockers and high levels of alcohol use, cholesterol, and forced expiratory volume in 1 second were associated with a reduced risk of AF. Conclusions The incidence of AF in older adults may be higher than estimated by previous population studies. Left atrial size appears to be an important risk factor, and the control of blood pressure and glucose may be important in preventing the development of AF.