Transthoracic electrical cardioversion using a monophasic waveform is the most common method converting persistent atrial fibrillation into sinus rhythm. Recently, cardioversion with a new biphasic waveform has shown promising results for treatment of atrial fibrillation. We undertook a randomized prospective trial comparing the efficacy and safety of the two waveforms for ambulatory cardioversion of atrial fibrillation. A total of 118 consecutive patients (mean age 62 years [SD 11]) presenting with persistent atrial fibrillation (mean duration 8 months [SD 11]) for ambulatory electrical cardioversion were randomized to receive either monophasic (n = 57) or biphasic shocks (n = 61). We used a standardized step-up protocol with increasing shock energies (100-360 joules) in either group. In all patients an anterior-posterior shock electrode position was used. If sinus rhythm was not achieved with the third (360 joules) shock, cardioversion was repeated with the opposite waveform. The two groups did not differ in demographic or disease-related data. The success rate was 100% for the biphasic and 73.7% for the monophasic waveform (p < 0.001). Biphasic patients required fewer shocks (1.5 versus 2.9) and a lower mean cumulative energy (203 versus 570 joules) (p < 0.001). Twelve out of 15 unsuccessfully treated monophasic patients were converted with biphasic shocks. The success rate for all 118 patients was 97.5%. No major acute complications were observed. For ambulatory transthoracic cardioversion of persistent atrial fibrillation biphasic shocks are of greater efficacy and require less energy than monophasic shocks. The procedure can be performed ambulatory and is safe regardless of shock waveform used.