In this paper, a new decomposition method of the interference EMG signal using MAMDF filtering and digital signal processor. The efficient software and hardware signal processing techniques are employed. The MAMDF filter is employed in order to estimate the presence and likely location of the respective templates which may include in the observed mixture, and high-resolution waveform alignment is employed in order to provide the optimal combination set and time delays of the selected templates. The TMS320C25 digital signal processor chip is employed in order to execute the intensive calculation part of the software. The method is verified through a simulation with real templates which are obtain ed from needle EMG. As a result, the proposed method provides an overall speed improvement of 32-40 times.