There is a challenging problem during camera calibration with fisheye lens using the traditional checkerboard corner detection method, and leads to low calibration precision of camera. So, an iterative correction checkerboard corner detection method for large distortion images is proposed. The main idea of this method is step by step iterative estimation and optimization of camera parameters through several shots. The camera initial parameters are obtained using the images from the remote-center small-distortion region. The distortion compensation of near-distance large distortion is realized through space and coordinate transformation and pixel interpolation gray. On this basis, the corner points of image edges can be detected, and the corner coordinates corresponding to the near-distance large-distortion images are calculated based on the parameter mapping relationship. Simulation and real image experimental results show that the proposed method is simple to be implemented, and improves the quantity and quality of corner detection effectively, which can satisfy the practical application requirements.