To extract the original features of a image exactly, a new image fusion method based on the curvelet transform is proposed. A series of original images are decomposed by the curvelet transform， and then a region image is obtained by an inverse curvelettransform and a domain processing. For the edges of high frequency areas, the edge distribution image is obtained by the activity of every area in different images and the region image is gotten by an interpolation. Finally, the image fusion is accomplished by Gaussian distribution sums for high frequency coefficients and by mean values for low frequency coefficients in the transform domain. A image fusion experiment is undertaken, and the experiment results indicate that the proposed method can obtain a better fusion image with high contrast, clear edges and more closed to a reference image as compared with that of conventional wavelet transform and pixelbased curvelet transform methods.