Miniature multi-rotor UAV-based synthetic aperture radar (SAR) systems exhibit slow flight speeds and are susceptible to airflow disturbances. This results in nonlinear radar trajectories during the synthetic aperture time. Furthermore, due to constraints in weight and volume, such systems often lack high-precision motion sensors, making SAR imaging exceptionally challenging. To address these issues, this paper proposes a SAR imaging processing method based on a two-stage compensation strategy. The algorithm first coarsely estimates the UAV’s velocity using the Map-Drift (MD) algorithm in the data domain. This coarse velocity estimate is then used as a reference to process the echo data via the Polar Format Algorithm, generating a coarsely focused image of the scene. Finally, a two-dimensional autofocus algorithm incorporating prior information is applied to refocus the coarse image, yielding a precisely focused image of the target. Extensive experiments with real-world data demonstrate that the proposed method achieves robust and high-precision imaging without motion sensor assistance. It significantly reduces the dependency on high-precision motion sensors for high-resolution SAR imaging. This approach holds considerable value for practical engineering applications.