To meet the need of object detection problem in images, a circle detection algorithm based on conformal geometric algebra and Radon transform is proposed. From a new point of view, the probability and simplicity for using the concept of conformal geometric algebra to solve circle detection problem is discussed in detail. Circles in Euclidean space can be represented as vectors by the conformal geometric algebra theory, a circle may be implemented as a plane in conformal space, and the plane in conformal space can be detected by 3D radon transform. This algorithm can detect circles in images effectively. The experimental results show that the proposed algorithm is compact and explicit, multi-circles in images can be detected by using this algorithm, and the integration of algorithm has an important significance in the improvement of efficiency.