The invention relates to a space-borne synthetic aperture interferometer radar image registration method with feature point Voronoi diagram optimization. The method is mainly characterized in that the deployment strategy of a matching reference point is novel, the attribute information of statistical information, thematic information, topological information, metric information and the like of a feature point is fully considered. The method comprises the steps of (1) obtaining the space-borne InSAR complex image pair of a measured area, (2) calculating the energy image of one complex image in the image pair and carrying out filtering processing, (3) using a Forstner operator to extract a feature point in the filtered energy image, (4) constructing a Voronoi diagram for the extracted feature point and optimizing the feature point by using the Voronoi diagram, (5) using the optimized feature point and adopting a method related to maximum spectrum to search a same name point, (6) calculating a model parameter between two images according to the searched same name point pair, and (7) unifying the two images in the same image coordinate system by using the calculated model parameter. According to the method, the same name point search success rate and related coefficient value are raised, thus the optimized feature point is more reasonable in the distribution, the calculation precision of the model coefficient is raised, and thus the registration precision of the image pair is raised.