In this paper, an approach is presented which removes the Electrocardiographic (ECG) interference from the Electomyographic (EMG) signal. When the EMG signal is taken from some specific muscles, it often gets contaminated with the noise and ECG signal. The method of non-stationary harmonic modelling effectively suppresses the noise and ECG signal and preserves the EMG signal. The major variations in the ECG signal are related to heart rate and QRS complex variability. These variations account for the continuous amplitude and frequency changes in the signal. Non-stationary harmonic modeling effectively captures these variations and produces the output as pure EMG signal. The proposed method outperforms the filter methods in terms of suppressing the noise component and preserving the EMG signal.