MODIS image is a kind of new and important data for global change research. After analyzing in the depth causes of MODIS data distortion

an optimal method to remove the geometric distortion is developed. The longitude and latitude coordinates

obtained by MODIS sensor

are used to remove the distortion of the 1km resolution MODIS data; For 250m and 500m resolution MODIS data

the same resolution coordinates are calculated by the spline interpolation

then the geometric distortion is removed. Because of the spatial irregularity of MODIS 1B data

a method with forward gridding approach and the inverse gridding approach is utilized to allocate a pixel’s position of the geometrically corrected data in the original image. By computing the overlap degree of the current position

the number of the pixels used to calculate the value of the output pixels

the search window sizes and the accurate position of the windows are decided. The value of output pixels is calculated by the distance inverse weight method. All optimal approaches above will improve the processed image’s quality and the processing efficiency. The software based on the above approaches is developed in the Visual C ++6.0 environment. From the result

the conclusion that the approaches are practical and reasonable can be made.