This paper proposes an improved wavelet coding for image de-noising in a super-resolution perspective. Here we present an improved version of wavelet based image de-nosing technique known as cycle-spinning to avoid the effect of Pseudo-Gibbs phenomena. With the thought of cyclic shift, this paper considers the relationship among each step of the whole reconstruction process, investigates the involved inverse translation and the average operation, and proposes two different image reconstruction methods. The experimental results indicate that these methods can eliminate the Pseudo-Gibbs phenomenon, and get better visual effects subjectively and higher PSNR value objectively.