The estimation of parameters under aff ine transformation plays a significant role in computer vision and has been foun d application in many fields.Images of an object taken from different viewpoints often suffer from perspective distortions.It is well known that an affine trans formation can be used to approximate the perspective transformation between two views.To align a pair of affine-related images better,we need to estimate the affine parameters between the two images.This pap er presents a method based on the central projection to estimate the parameters under affine t ransformation.The algorithms of estimating parameters under affine transformation can be generally classified into two categories:region-based methods and contour-based methods.The contour-based methods are easy to implement,but generally they can not deal with mutil-part object.However,region -based methods make computation heavier and sensitive to noise.This paper will combine contour-bas ed methods with region-based methods by central projection.Firstly,a closed curve is derived fr om the object using central projection and the closed curve can ensure the affine transformation rel ation with invariant property.Then,the closed curve is parameterized.Finally,the affine transformatio n parameters are estimated by the parameterized closed curve.The experiment results show that the proposed method can be used for the estimation of affine transformation parameters of mutil-part objec t and has a better parameter estimation effect.