Summary: MR Diffusion Weighed Image (DWI) is one of many functional magnetic resonance imaging (fMRI) techniques, and could provide complicated spatial and structural information about the tissue. Aiming at segmentation MR diffusion weighed image for real-time application with numerical stability constraints and high efficiency, a method based on the minimization algorithm is developed. Our approach is based on the image segmentation tasks into a global minimization method. The minimization algorithm minimization the energy, avoid the drawback in the level set approach and easy to implement, allows us a fast minimization of the active contour. Experimental results show that the effectiveness for image segmentation with our method is preferable.