Abstract: ‍ ‍Terrain observation by progressive scans （TOPS） is a working mode of spaceborne SAR that has a wide swath. Because of this characteristic， spaceborne TOPS SAR finds wide application in the military and civilian fields. Under wide swath conditions， there tends to be a relatively wide range of target types within the TOPS SAR scanning scene. The azimuth resolution requirements of the targets increase with the increase in the number of target types. However， the azimuth resolution of conventional TOPS SAR is single， which does not accommodate the varying resolution requirements of different target types over a wide scanning range. Spaceborne TOPS SAR with azimuth-varying speed scanning addresses this limitation by adjusting the scanning rate of the SAR antenna beam. This adjustment allows targets at different azimuth positions to be scanned for varying durations， enabling the acquisition of the appropriate information about the targets on the basis of their azimuth positions. Consequently， spaceborne TOPS SAR with azimuth-varying speed scanning can achieve multiple azimuth resolutions within the same scanning area to meet the azimuth resolution requirements of a diverse array of target types simultaneously. However， the conventional signal processing method is no longer adequate for spaceborne TOPS SAR with azimuth-varying speed scanning. Therefore， this study proposes a novel signal processing method that is suitable for spaceborne TOPS SAR with azimuth-varying speed scanning， based on its signal characteristics. Initially Doppler centroid removal processing and azimuth zero padding processing to eliminate the Doppler aliasing phenomenon as much as possible. In the second step， the total azimuth Doppler bandwidth of the signal is made less than the pulse repetition frequency （PRF） of the SAR system via deramping processing， upsampling processing， and other forms of processing. Thus， the Doppler aliasing phenomenon of the signal is completely eliminated. In the third step， the Doppler history of the signal prior to processing is reconstructed by multiplying the original Doppler history recovery function. Next， the range cell migration correction and range compression are performed. Finally， the instantaneous signal phase shift function and scaled Fourier transform （SCFT） are used to mitigate the influence of the distance variation from the beam rotation center to the satellite on the azimuth range sampling interval. Moreover， the azimuth compression is performed to generate the SAR image. Through point target simulation experiments， scene distributed target simulation experiments， and a comparison of the imaging performance indicators of the targets， the validity of the target imaging results and effectiveness of the proposed signal processing method for spaceborne TOPS SAR with azimuth-varying speed scanning are verified. The proposed method effectively addresses the Doppler aliasing problem of the signal， can accurately image a target， and achieves different azimuth resolutions within the same swath. Simultaneously， the proposed method meets the azimuth resolution requirements of various target types over a wide scanning range and provides an ideal and convenient solution for application scenarios that require different azimuth resolutions.