Doppler techniques used in blood flow velocity estimation rely on the assumption that the phase of the envelope-detected echo signals changes linearly with depth. However, the random spatial distribution of scatterers makes it inadequate. Correlation searching techniques utilize time shifts to remedy this drawback, but the probability of the correct maximum detection is also low under certain measurement conditions. In this study, a more sensitive and simpler means of velocity estimation has been proposed. This method tracks a segment of the echo footprints by means of the least-sum-squared difference from the consecutive RF echo signals. Since a squared difference measure is used to extract the moving information, the detection results are sensitive to the locally large differences, no matter how large the amplitudes are here, and the probability of finding a correct match by this method is highly increased. A detailed computer simulation evaluated the interdependencies of the window size, the pulse duration, the velocity searching scope, and the aliasing problem on the estimation of the blood flow velocity. The results are proved to be more accurate and effective than those of the conventional techniques.