In this paper, complexity measure based on Jensen-Shannon Divergence was used to compute complexity of the ECG signals, which include normal sinus rhythm, atrial premature contraction (APC) and sinus bradycardia (SBR) signals from the MIT-BIH standard database. The results show that three kinds of signals have different complexity measures. Normal sinus rhythm has the highest complexity, followed by sinus bradycardia (SBR) signals, and atrial premature contraction (APC) has the lowest complexity. It has good reference for clinical detecting and diagnosing with atrial premature contraction (APC) and sinus bradycardia (SBR) signals.