Abstract: A coordinate inverse transform (CIT) calculates the image-space coordinates of a corresponding ground point, and is substantially important for the processing of linear array pushbroom images. One of the imaging characteristics of a TDI CCD pushbroom sensor is that the scanline integration time fluctuates. CIT is performed on TDI CCD pushbroom images, based on its collinearity equations; the process will iterate several times, and even then sometimes, no convergent results are available. Since the fluctuation of integration time also affects the alternative accuracy of generic imaging models such as the rational function model (RFM), the accuracy of CIT based on the RFM of TDI CCD pushbroom images can be hardly guaranteed. Hence, this paper proposes a novel method for a CIT of TDI CCD pushbroom images based on virtual image scanline, solving the problem of low efficiency and accuracy when calculating image-space coordinates. Experiments tested on the TDI CCD pushbroom images from a domestic resource satellite demonstrate the correctness and feasibility of our method.