A novel de-noising method, called matching pursuit method, for improving the signal-to-noise ratio (SNR) of Doppler ultrasound blood flow signals is proposed. Using this method, the Doppler ultrasound signal is first decomposed into a linear expansion of waveforms, called time-frequency atoms, which are selected from a redundant dictionary named Gabor functions. Then a decay parameter-based algorithm is employed to determine the decomposition times. Finally, the de-noised Doppler signal is reconstructed using the selected components. The SNR improvements and the maximum frequency estimation precision with simulated Doppler blood flow signals have been used to evaluate a performance comparison based on the wavelet, the wavelet packets and the matching pursuit de-noising algorithms. From the simulation and clinical experiment results, it is concluded that the performance of the matching pursuit approach is the best for the Doppler ultrasound signal de-noising.