In dental imaging, Cone-beam computed tomography(CBCT) is widely used due to its high resolution, low radiation dose, and cost. However, motion artifact is a frequent issue due to difficulty in maintaining a stable posture during long scanning time. Therefore, we propose a method of reducing the motion artifact in CBCT imaging using D-Neural Radiance Fields(NeRF) generating a novel synthesized images in dynamic scenes. The experimental results indicate our model significantly outperforms other reconstruction tasks with three motion.