The invention discloses an ultrasound contrast image segmentation method based on a statistical partial differential model. The method comprises a preprocessing step and an online segmentation step. At the preprocessing step, a to-be-segmented image and an initial contour are obtained; and at the online segmentation step, an energy function is obtained by using the to-be-segmented image and the initial contour as inputs, an edge stop function is introduced into the energy function to constrain the length item and an LoG image energy function is added, numerical solution is carried out on the energy function, and target edge capturing and segmentation are carried out on a single or a plurality of different ROI images quickly. The ROI in an ultrasound contrast image is segmented; a doctor isassisted in obtaining time strength curves of masses in a lesion region and same level reference regions like parenchymal organs including the liver and the kidney and the like and superficial organsincluding the breast and the thyroid gland and the like by observing the segmentation result of the whole sequence, so that a quantitative index is obtained by the analysis.