In northern hemisphere of the Earth, affected by factors such as the orbit characteristics of Sun-synchronous satellite, source-observer position, solar azimuth, solar elevation angle and topographic relief while the sensors are imaging, topographic features of satellite images have false topographic perception phenomenon (FTPP). Although the traditional SRM-based HIS (HSV, HLS) transformation method can effectively remedy FTPP of remote sensing images, there exists a problem of serious distortion of spectral information. Firstly, this paper proposed a new SRM weighted by surface roughness FTPP correction method, and made effective FTPP correction for GeoEye-1 satellite RGB color images at both 2.5-meter and 10-meter spatial resolution. Then the SRM weighted by surface roughness method first converted a remote sensing image from RGB color space into HSV color space to obtain the color value data (V) of image. Besides, this paper took linear standardized surface roughness( )as weights of simulated northwest illumination shaded relief data and took 1- as weights of color value component (V) to sum up a new color value component (V’). Lastly, based on the new color value (V’), it transfered the image back from HVS to RGB format to achieve remote sensing image FTPP correction. Compared with SRM based HSV conversion FTPP correction method, this method takes into account of the influence of terrain factors on FTPP, that is, making effective FTPP correction in rugged areas while retaining most spectral information in flat terrains, namely, area of smaller surface roughness of the remote sensing images. Lastly, this paper proposed taking the variance of the color value (V) component used when HSV was retransformed to RGB color space and original color value (V) component value as the metric for a quantitative evaluation of the amount of image spectral information distortion after FTPP correction, and made a quantitative analysis of image spectral information distortion resulting from the two methods of FTPP correction – the SRM weighted by surface roughness and the traditional SRM based HSV transformation method. The results show: after FTPP correction for GeoEye satellite images at 2.5-meter spatial resolution by the two methods, the average distortion of spectral information were: -0.057715 and -0.110918 respectively; and relative standard deviation were 0.131862 and 0.247324 respectively. Image spectral distortion resulting from the use of SRM weighted by surface roughness method was reduced by 46.684511% compared with that of the latter. After FTPP correction for GeoEye satellite images at 10-meter spatial resolution by the two methods, the average distortion of spectral information were -0.066658 and -0.105606 respectively; and relative standard deviation were 0.153065 and 0.238166 respectively. Image spectral distortion resulting from the use of SRM-base weighted by surface roughness method was reduced by 35.731809% compared with that of the latter. Quantitative spectral information distortion results show that compared with the traditional SRM based HSV transformation method, the use of SRM weighted by surface roughness FTPP correction method retains more image spectral information and also has certain universality at different spatial scales.