Abstract: This paper describes a preprocessing mask technique based statistical mixture components segmentation method for extracting blood vessels from brain magnetic resonance angiography (MRA) dataset. The voxels whose intensity is high in the dataset belong to blood vessels or brain skulls, which may bias the adjustment of the blood vessels. Maximum intensity projection (MIP) of the dataset in the Z axis direction was computed and segmented as a mask. The masked MRA dataset was segmented by a low threshold and the remanent voxels were modeled by one normal distribution and one uniform distribution. The parameters were estimated by Expectation-Maximization (EM) algorithm. The results show that this method is feasible for vessel extraction from MRA dataset.