Abstract: A multivariate statistical method was proposed to correct ring artifacts in industrial CT images. First, the grayscale histogram statistics and maximum interclass variance were used to classify the density in CT images. Then, the coordinate system was converted to the polar one and the abscissa pixel values were counted in the polar coordinate image and the artifacts of the gray scale amplitude were extracted at different radii. Finally, the correction was performed by subtracting artifacts from the original CT image. The CT images of different structures with ring artifacts were calibrated by using 6 MeV linear array detector high energy industrial CT system. The experimental results show that this method can effectively remove CT images from ring artifacts. The image details and resolution were well kept and the processing speed was significantly improved.