The frame transfer CCD with high sensitivity, high resolution and wide spectral response is the main image sensor of the satellite-borne polarization camera. The image fringe blurring phenomenon caused by the inherent structure of the frame transfer area array CCD is called smearing, which will reduce the imaging quality and spectral measurement and then affect the accuracy of remote sensing image product parameter inversion. Therefore, research on the evaluation standard and correction method of the tailing degree is of great significance to realize the high-precision measurement of satellite-borne polarization camera. Based on the working principle of the directional polarization camera, which assembles the frame transfer CCD, this paper analyzes the mechanism of smearing and introduces the dark line correction model by using the dark line of frame transfer CCD. According to the characteristics of the dark line correction model, this paper analyzes the light leakage problem of some dark line pixels adjacent to the photosensitive area due to the point diffusion effect, and it is concluded that different dark lines have an effect on the smearing correction effect. To avoid the problem of inaccurate correction of smearing caused by light leakage dark lines, this paper proposes a method to compare the degree of image smearing to select the appropriate number of dark lines. That is, the degree of smearing of the image corrected by different numbers of dark lines is evaluated by appropriate criteria. After comparison, the image with the lowest smearing degree after correction is selected as the image with the best correction effect, and the corresponding image is the most appropriate number of dark lines. To find suitable evaluation criteria, the smearing simulation model of the integrating sphere imaged by the directional polarization camera is established by using MATLAB. Combined with the smearing simulation model, the gray standard deviation and average gradient of the smearing area are used as the evaluation criteria of the image smearing degree. The smearing stripes with different brightnesses of the integrating sphere are obtained by using the simulation model, and then the gray standard deviation and average gradient in the smearing area are calculated. Then, the experiment of imaging different luminance integrating spheres by directional polarization camera is carried out, the same brightness as the simulation integrating sphere is set, and the experimental images under different luminances of integrating sphere are obtained. The gray standard deviation and average gradient of the smearing area in the experimental image are calculated to verify the simulation data. The results show that the simulation data are in good agreement with the experimental data, which verifies the effectiveness of the evaluation criteria. Finally, to solve the problem of inaccurate smearing correction caused by light leakage dark lines, a correction smearing optimization algorithm is proposed according to the evaluation criteria of the gray standard deviation and average gradient, which is verified by integrating sphere imaging at the edge of the field of view of the directional polarization camera. The algorithm can calculate and compare the gray standard deviation and average gradient of the smearing image corrected by different dark lines in each band. The experimental results show that the correction smearing optimization algorithm can adaptively determine the most appropriate number of dark lines needed to correct smearing in the current experimental environment. Using this algorithm, the gray standard deviation percentage of the image smearing after the correction of each channel of the directional polarization camera is reduced by 50.3%~89.9%, and the average gradient percentage is reduced by 65.2%~74.0%. The research results provide a reference for the detection and correction of image smears by satellite-borne polarizing cameras.