Annular field aberrations of a three-reflection concentric system, which are composed of two spherical mirrors, are analyzed. An annular field with a high level of aberration correction exists near the position where the principal ray is perpendicular to the object-image plane. Aberrations are determined by the object height and aperture angle. In this letter, the general expression of the system aberration is derived using the geometric method, and the non-approximate design method is proposed to calculate the radii of the annular fields that have minimum aberrations under different aperture angles. The closer to 0.5 (the ratio of the radius of convex mirror to the radius of concave mirror) is, the smaller the system aberration is. The examples analyzed by LABVIEW indicate that the annular field designed by the proposed method has the smallest aberration in a given system. Concentric optical system is constituted by a series of refracting or reflecting spherical surfaces that have the same center and separated by homogeneous media. Any plane containing the optical axis is symmetrical, and any ray traversing through the spherical center is the axis of the rotational symmetry of the system. Therefore, this system can eliminate high-order aberrations, and consequently, it can be widely applied in a remote-sensing system [1,2] as well as in the large-scale integrated circuit manufacturing industry [3] . Generally, the concentric optical system can be divided into systems with a single mirror [4,5] , two mirrors, and three mirrors. In 1973, Offner proposed a kind of two-mirror three-reflection concentric system [6,7] . The author pointed out that an annular field was present in this concentric system, which could be considered to have zero aberration because the fifth-order aberration could be balanced by the third-order aberration [8] . In 1987, Charls [9] analyzed the concentric optical system in detail and proved that the concentric optical system had the following characteristics. Firstly, the product nd for any ray that traces such a system will maintain the same value throughout, where n is the refractive index of the space medium and d is the perpendicular distance of the ray from the spherical center of the space. Secondly, if the object and the image are in the air and the incident ray emerges as parallel, then the system will perform ideal imaging at a point located at the joint of the object and the system center. Thirdly, the distance of the incident and emergent rays from the center will be in the ratio of the object to the image space refractive indices. Fourthly, for a system consisting only of refracting surfaces or a system with an even number of reflections, the system will produce virtual images. The rays will be positioned on the same side of the system center. For a system with an odd number of reflections, real images will be produced, and the incident and reflected rays will be located on the opposite sides of the center. This is true for the Dyson system [4] and the Offner system [8] . By car...