Abstract: An index for blur assessment was proposed based on gradient analysis of gray gradient cooccurrence matrix in wavelet domain.Based on the natural scene statistical feature of remote sensing image, the means of different sensitivity for noise and blur in different wavelet subbands, an antinoise blur assessment index was designed for remote sensing images.The experiment shows that the proposed index has high coherence with real blur intensity and can make an effective and correct evaluation of blur level for remote sensing images.