Space-based segmented telescopes are susceptible to mirror misalignments because of temperature and mechanical perturbations in orbit. Monitoring the misalignment status of each mirror is a prerequisite to aligning the telescope promptly and maintaining image quality. In this paper, an online monitoring method based on an improved vision transformer network is proposed. Two new modules have been added to the method: the grid-aligned patch-splitting (GAPS) module and the field linearization eliminator (FLE) module. The former modifies the patch-splitting method based on the characteristics of point spread function (PSF) to improve the accuracy. The latter is used to eliminate the effect of field position for arbitrary field-of-view wavefront monitoring. The method predicts the misalignment state of the telescope through a PSF (stellar image) for an arbitrary field of view, realizing monitoring of the telescope without hindering the observing process and increasing the complexity of the system. Simulations and experiments are performed to verify the effectiveness and practicality of the method.