Abstract: ‍ ‍In this paper， the image feature analysis of the moving target in bistatic synthetic aperture radar （bi-SAR） under arbitrary configuration has been studied in detail based on phase error in the wavenumber domain. Different from the conventional approach where phase error analysis is performed in the phase history domain， in this paper， we provided a detailed analysis of the 2-D phase error in the PFA wavenumber domain. The analysis method is also applicable to other algorithms such as the BP algorithm. The general process of SAR imaging algorithm involved converting the signal from the phase history domain to the wavenumber domain， and then transforming it to the image domain. However， the process of converting signal from the phase history domain to the wavenumber domain may introduce new phase errors. Therefore， analyzing the characteristics of phase error in the phase history domain may not be entirely accurate. As a result， there may be some errors in the imaging results predicted by the characteristics observed in the wavenumber domain. Since there is a direct Fourier transform relationship between the wavenumber and image domains， and the properties of the Fourier transform are already well-known， we can accurately predict the imaging characteristics of moving target in the PFA imagery by using the Fourier transform relationship between the two domains and considering the phase error characteristics observed in the wavenumber domain. This article presents a significant finding that the 2-D phase error in the wavenumber domain can lead to a 2-D position shift and defocusing of targets in the image domain， both occurring along a skewed straight line. This means that， regardless of the direction of the moving target motion， the image of the moving target in the image domain will consistently shift in a specific line. Alternatively， by implementing an appropriate coordinate rotation， the dimension of the phase error can be reduced from 2-D to 1-D. This simplification of the phase error dimension can， in turn， minimize the complexity and difficulty of relocating and refocusing on the target in the image domain. Finally， the simulation experiments are carried out to verify the predicted signatures of the moving target in PFA imagery.