In Interferometric Inverse Synthetic Aperture Radar (InISAR) three-Dimensional (3D) imaging, Inverse Synthetic Aperture Radar (ISAR) image registration is a key process. It arranges similar scatterers in different ISAR images on the same pixel cell, which completes the interferometric processing for similar scatterers to obtain their 3D coordinates. This paper analyses the ISAR image mismatch caused by the difference of the wave path between the scatterer and different radars and then proposes a new ISAR image registration method that consists of three steps, based on wave path difference compensation. First, chirp Fourier transform is used to estimate the rotational angular velocity of the target. Second, the compensation phase is constructed according to the rotational angular velocity to eliminate the wave path difference between scatterer and different radars. Finally, two-Dimensional (2D) Fourier transform can yield ISAR images without mismatch. The proposed method can directly achieve ISAR image registration in an echo field. After image registration based on this method, the scatterers match-up well. But after image registration based on the correlation method, the scatterers have a mismatch of about one pixel cell. Compared to the correlation method, this method can achieve image registration more efficiently. Finally, the 3D imaging error (0.3034) based on this method is smaller than that (45.8529) of the correlation method.