In general electromyography (EMG) is used to find the activity of muscles. When we want to extract accurate information, it is required to record clean and undistorted electromyography (EMG) signals. Generally, when EMG is recorded of some specific muscles, it is often contaminated by ECG signal, hereby significantly increasing the power of EMG signal. This can hardly be avoided; therefore, to extract valid information it is necessary to process EMG signal to remove ECG signal. Here we model such that variations of the ECG in terms of amplitude and frequency time are evaluated by using the Heart rate and QRS complex; the respective variations are simultaneously captured by a set of third-order constant-coefficient polynomials modulating a stationary harmonic basis in the analysis window.