An edge point matching method of SAR image based on the joint similarity is presented. First, the matching similarity criterion and the joint similarity——SSJF(Square Summation Joint Feature) were constructed, based on the grad and direction of each edge point in images. Next, a modified ROEWA edge detector was proposed to get the edge intensity and edge direction with the eight directional templates. Then, parameters of the transform model between the matching SAR images were calculated with the modified GA, which is used to obtain the global optimum solution of the similarity. Finally, the performance of the method was validated with SAR image registration experiments.