Traditional image detection methods do not obtain the brightness, chroma and saturation information of remote sensing images, and the Performance of image saliency region is poor. Therefore, this paper proposes a detection algorithm of remote sensing image saliency region based on wavelet denoising.The remote sensing image is transformed into one dimensional image by orthogonal wavelet transform, the edge information of remote sensing image is obtained by binary wavelet transform, and the edge noise is removed according to the new threshold function. The brightness, chroma and Saturation information of remote sensing images were calculated according to the IHS transform (Intense-Hue-saturation) method, and the high and low frequency coefficient vectors of remote sensing images were calculated by discrete wavelet transform. After obtaining the clustering data of low frequency coefficients through fuzzy C-means clustering, significance factors were used to complete the detection of the significance region of remote sensing images.The experimental results show that the proposed method can effectively extract image edge information, has a strong ability to remove noise, and the detected image saliency region is well arranged, with high contrast and good detection effect.