Abstract: Arc SAR is a radar imaging technology that can scan 360° repeatedly in a short time. It has the advantages of short revisit period and wide observation range, and has a wide application prospect in slope detection, landslide and other scenes. In Arc SAR mode, the radar moves in a circle along with the robotic arm to form synthetic aperture. Because its motion path is not a straight line, the traditional linear guide imaging method cannot be directly applied. A new imaging algorithm needs to be designed for Arc SAR, but the existing imaging methods cannot balance accuracy and complexity. In the range Doppler algorithm, because Taylor expansion is used to approximate the oblique distance, the image quality is not good, or segmentation strategy is needed for imaging. A new range Doppler algorithm is proposed. The approximate slant range expression is obtained by using the vector relationship between target and radar, and then the signal expression in range Doppler domain is obtained. Based on this expression, the corresponding phase error compensation method is proposed to complete the image focusing. Finally, the effectiveness of the proposed algorithm is verified by point target simulation.