Abstract: A novel method that can be used for improving line-scan camera calibration according to its features was proposed. Firstly, the designed planar pattern and multiple exposures of line-scan camera were used to produce a virtual stereo calibration pattern. Secondly, a mathematical model was analyzed and established to seize the relation between the coordinates of calibration points on the pattern and those of their images. Then, the technique of least mean square and iterative optimization were employed to solve both the over-determined equations for multiple feature pointsand nonlinear parameters of lens. The impact to the calibration result that was caused by the non-parallel nature between the camera and calibration pattern was also analyzed. Finally, both the analysis and experimental results reveal that this proposed calibration method is not only convenient and flexible, but the calibration points are also independent of calibration pattern and that the condition of parallelism is not required. All these features make the devised method accurate and practical.