Abstarct: Image registration is a way to match two images of a scene taken at different times or taken using different sensors. One of the purposes for matching of two images captured from a scene using different type of sensors is to utilize information provided from different image sensors. Different geometric features and radiometric characteristics for the two type of images, as well as the strong speckle noise in radar images, cause the registration of SAR and optical images becomes severe. In the proposed method we first denoise the SAR image, and then extract the straight lines of both images. After edge detection stright lines are extracted using the radon transform. The extracted lines using radon transform are much more accurate than the Hough transform and LSD. Each pair of straight lines intersect at a point. We characterize each intersect point using its voronoi cell. Voronoi cells divide the image into subregions where each cell surround an intersect point. By calculating the Hausdorff distance of the Voronoi cells and the eigenvector method, we match SAR and optical images. The proposed method has been able to properly match the intersect points of two images. This results a proper matching of SAR and optical images.