Open AccessThe ECG recording is highly vulnerable to various kind of noises from different sources, such as electrocardiogram (EMG), power supply hum (50Hz or 60 Hz), measuring devices such as amplifiers, ADC etc. Hence it is very difficult and challenging to interpret and analyze raw ECG data for medical applications. A number of techniques are available to deal with these types of noises efficiently both during recording and pre processing of ECG data. In this paper, four different wavelet threshold denoising techniques are proposed to deal with the issue of noises in ECG recording. Analysis and their performances have been evaluated in terms of SNR and RMSE. The standard MITBIH arrhythmia data from physionet is used for the purpose. The procedure is implemented in MATLAB environment.