The risk, correlates, and consequences of incident atrial fibrillation (AF) in patients with chronic coronary artery disease (CAD) are largely unknown. We analyzed incident AF during a 3-year follow-up in 5031 CAD outpatients included in the prospective multicenter CARDIONOR registry and with no history of AF at baseline. Incident AF occurred in 266 patients (3-year cumulative incidence: 4.7% (95% confidence interval (CI): 4.1 to 5.3)). Incident AF was diagnosed during cardiology outpatient visits in 177 (66.5%) patients, 87 of whom were asymptomatic. Of note, 46 (17.3%) patients were diagnosed at time of hospitalization for heart failure, and a few patients ( = 5) at the time of ischemic stroke. Five variables were independently associated with incident AF: older age ( < 0.0001), heart failure ( = 0.003), lower left ventricle ejection fraction ( = 0.008), history of hypertension ( = 0.010), and diabetes mellitus ( = 0.033). Anticoagulant therapy was used in 245 (92%) patients and was associated with an antiplatelet drug in half ( = 122). Incident AF was a powerful predictor of all-cause (adjusted hazard ratio: 2.04; 95% CI: 1.47 to 2.83; < 0.0001) and cardiovascular mortality (adjusted hazard ratio: 2.88; 95% CI: 1.88 to 4.43; < 0.0001). In CAD outpatients, real-life incident AF occurs at a stable rate of 1.6% annually and is frequently diagnosed in asymptomatic patients during cardiology outpatient visits. Anticoagulation is used in most cases, often combined with antiplatelet therapy. Incident AF is associated with increased mortality.