Blood flow plays an important role in clinical diagnosis and treatment, and noninvasive measurement of blood flow is the expectation of both doctors and patients. In the present work, the authors employed near-infrared diffuse correlation spectroscopy (DCS) method to measure blood flow noninvasively. The relationship between speckle intensity fluctuations of the scattering light and moving blood cells in tissue was theoretically analyzed. A theoretical model and blood flow measurement system were built. Blood flow was derived by calculating the electric field temporal autocorrelation function of speckle on tissue surface. Forearm blood flow was measured in healthy human subject during cuff inflation and deflation. Experimental results show that noninvasive blood flow measurement by this system is feasible. The results also suggest that this approach can provide blood flow information throughout the whole depth profile of the tissue.