Abstract: As an important technology in the direction of optical quantum science, single photon detection technology has been widely used in many fields such as autonomous driving, ground remote sensing and aerospace. The three-dimensional imaging target is often in motion, resulting in structural offset and registration error between image frames, which affects the clarity and accuracy of imaging. Aiming at this problem, we proposed a dynamic single-photon three-dimensional imaging method based on adaptive extended Kalman filter. Combined with the image quality assessment results, the covariance parameters were adaptively adjusted, and the matching features between images and the timing weighting strategy were fused to achieve the stable estimation of the target rotation state. The experimental simulation was verified based on the generated single-photon three-dimensional imaging data. The results show that the proposed method can effectively restore the target contour and structural details in complex scenes. Compared with the traditional method, the average PSNR (peak signal-to-noise ratio) is increased by 2.93 dB, and the SSIM (structural dissimilarity) is increased by 0.115, which verifies its imaging accuracy and robustness advantages under complex rotation conditions.