Aiming at the problem of complex range processing and strong azimuth variant in GEosynchronous Orbit SAR (GEO SAR) imaging processing, an imaging algorithm based on Chirp_z transform and the azimuth scaling is proposed. Firstly, the “stop-go-stop” phase error is considered in the echo model, and then the slant distance expression which is expanded using the Taylor series is given. After that, a series of inversion method is used to get the signal spectrum expression. In order to simplifying the imaging processing at range direction, the Chirp_z transform is used to correct the range walk. And then, an improved scale factor is used to correct the linear time-varying coefficient of signal expand included the first order, the second order and the three order. The phase error introduced by scaling operation is compensated in frequency domain. Finally, the simulation results show that the algorithm can realize geosynchronous orbit SAR imaging.