Multi-scale curvelet transform is a new extension to wavelet transform in two dimensions. The directionality feature of curvelet transform makes it a good choice for representation of curves and edges in the image. The second generation curvelet transform theory makes it understood and implemented more easily. In this paper, a novel algorithm is proposed to improve compression performance by using the second generation curvelet transform and SVM (Support Vector Machine) regression. The method is also compared with image compression method based on wavelet transform. Experimental results show that the compression performance of our method gains much improvement. Moreover, the algorithm works fairly well for declining block effect at higher compression ratios.