Introduction: Atrial fibrillation is the most commonly encountered arrhythmia in cardiological practice, its predominant form in developed countries is non-valvular atrial fibrillation which constantly increasing even in developing countries, further its thromboembolic complications this arrhythmias is associated with a higher risk of mortality, heart failure, and hospitalization. The aim of this work is to study the epidemiological profile of non-valvular AF in Morocco, to evaluate the therapeutic management and to investigate the part of the different risk factors for embolic events in our local context. Materials and methods: an observational study including 250 patients admitted for non-valvular atrial fibrillation was conducted in the cardiology department of the military training hospital Mohammed V in Morocco, all patients underwent electrocardiogram, echocardiography and biological determination of thyroid hormones and renal function their clinical characteristics as well as echocardiographic and therapeutic data were collected and analyzed. Results: The mean age of our patients was 63 years (+/- 12) with a sex ratio H / F of 2.1. Atrial fibrillation was permanent in 66% of patients, paroxysmal in 23% and persistent in 11% of cases. The mean diameter of the left atrium was 44 mm (+/- 7) with a mean surface area of 25 cm2 (+/- 7), the average ejection fraction was 52% (+/- 13). The most common symptoms were palpitations, dyspnoea and systemic embolism. The average CHA2DS2Vasc score was of 2.57, only 58% of our patients were on VKA at admission. Among the different embolic risk factors studied, only age diabetes and hypertension were significantly associated with the occurrence of a thromboembolic accident in our study. A strategy of rhythm control was adopted in 75 patients: 45 patients underwent cardioversion and 30 other underwent pulmonary veins isolation. Conclusion: the epidemiological profile and clinical features of non-valvular atrial fibrillation are different in our local context, an under-utilization of anticoagulant therapy in eligible patients is noted as well as a less frequent use of a rhythm control strategy compared to a rate control strategy.