Autonomous navigation technology has played an increasingly important role during modern planetary exploration missions. The image-based optical navigation method is one of the most promising solutions for autonomous navigation. Previous studies have mainly focused on image processing and navigation filtering algorithms. Classic ellipse fitting approaches rely on the least square method, which may fail in the case of short-arc ellipse fitting. This paper proposes an optimization method with a priori shape information to achieve a high-precision short-arc ellipse fitting. First, by querying the ephemeris and using the projection principle, the projected ellipse of the entire planet can be obtained. Second, the shape information of the ellipse that represents the useful priori information can be extracted using the least squares ellipse fitting algorithm. Third, the shape information is introduced into the standard elliptic equation, whose parameters are optimized to fit the planetary edge arc obtained by the navigation camera. Finally, simulations are conducted and the results show that the proposed method performs better in the accuracy of short-arc ellipse fitting compared to the traditional least squares method.