The treatment of atrial fibrillation (AF) remains one of the most difficult tasks of modern cardiology. In the past it was considered acceptable to base the pharmacological therapy of AF on an empirical approach. However, when making therapeutic decisions in the actual era of evidence based medicine it is important to favour the strategies that are grounded on data proving their efficacy and safety. Antiarrhythmic drugs in the truest sense of the word are used only for pharmacological cardioversion and prevention of AF recurrences. In the last years several studies were published that compared the strategy of maintaining the sinus rhythm, so-called rhythm-control strategy, with the strategy of controlling the heart rate during AF. As for the mortality impact, no superiority of rhythm-control strategy to rate-control strategy was shown. Although post-hoc analyses finally suggest that rhythm control represents a survival improving factor, these studies have significantly influenced the treatment of AF. At the present, we apparently do not dispose of any “ideal” antiarrhythmic drug that would be deprived of side and proarrhythmic effects, would affect with the same efficacy the atrial and ventricular arrhythmias and could be taken in longer time intervals. Even never class III antiarrhythmic drugs did not clearly demonstrate their superiority to classical agents and the questions of their efficacy and safety remain open. Therefore it is nowadays very important to carefully consider the need of restoring and maintaining the sinus rhythm. If cardioversion is decided, it is necessary to use the available drugs reasonably and in order to lower the arrhythmic burden, improve patient`s quality of life and not to endanger his life by the proarrhythmic effects of the chosen drug.