Single point active alignment method (SPAAM) has become the basic calibration method for optical-see-through head-mounted displays since its appearance. However, SPAAM is based on a simple static pinhole camera model that assumes a static relationship between the user’s eye and the HMD. Such theoretic defects lead to a limitation in calibration accuracy. We model the eye as a dynamic pinhole camera to account for the displacement of the eye during the calibration process. We use region-induced data enhancement (RIDE) to reduce the system error in the acquisition process. The experimental results prove that the proposed dynamic model performs better than the traditional static model, and the RIDE method can help users obtain a more accurate calibration result based on the dynamic model, which improves the accuracy significantly compared to the standard SPAAM.