Electrocardiography(ECG) plays a very important role in the diagnosis and treatment of heart disease, and R peak detection has become an important part of the ECG processing due to the Rwave amplitude with a large, easydetection feature. Correct detection of R peak is vital to the detection of heart rate and electrophysiological analysis. However, the ECG signal in the acquisition process will be subject to a certain degree of interference, such as frequency interference, baseline drift and other issues,so common R peak threshold algorithm can not detect lower baseline ECG waveform, resulting undetected. To solve this problem, this paper presents a R peak improved threshold extraction algorithm by setting thresholds and reexamining the mechanism for each cardiac cycle to improve the accuracy of the threshold algorithm. Simulation results show that the method is much simpler and more effective, and can realtime processing, which significantly improves the accuracy of the R peak detection and meets the requirements of clinical applications.