Because of cloud in remote sensing images, the information of earth objects could not been read. A new method for removing cloud had been projected based on wavelet analysis. Because of the cloud in remote sensing images was in the part of low frequency, the new method enhanced the sence information by various approximation coefficients and detail coefficients with various weight after wavelet transform. The criterions of entropy and fidelity had been applied to evaluate the quality of removed cloud images. The result shown the cloud in remote sensing images was removed by wavelet transform. First, the effect of removal cloud was good when the detail coefficients were multiplied weight and the information of sence was enhanced for decomposition order over 6. Second, the fidelty of the good enhanced images was 0.87±0.03 and 0.89±0.04 for thick cloud and for thin cloud, respectively. Finally, the color could be restored for removal cloud images by Brovery transform. In particular, Brovery transform was suit for removal thin cloud in remote sensing images.