The control of vasospasm is important in postoperative management after the occurrence of subarachnoid hemorrhage. Transcranial Doppler ultrasound (TCD) is a diagnostic safety test that measures the blood flow in the cranium; however, TCD has low reliability because of the high-intensity interference returned from the cranium. In this study, we employed two techniques to suppress interferences. First, we used a moving target indicator (MTI) filter that suppresses echoes from undesired stationary targets. Second, we applied spatial domain interferometry with the Capon method to the signal after MTI filtering. The desired signal from red blood cells is supposed to have a low correlation with the interferences from undesired stationary targets. We thus averaged the covariance matrix to range direction without averaging sub-matrices in the diagonal direction. In a simulation study, we succeeded in estimating blood flow velocity with an average estimation velocity error of 0.2 m/s, where the desired signal intensity from red blood cells is 60 dB lower than the cranium interference intensity. The center frequency of the pulse is 2.0 MHz and the pulse length is 5 μs. These results indicate the high potential of the proposed method for improving the estimation accuracy of blood flow velocity in the cranium.