MagicDance is a novel and effective method for generating realistic human videos, enabling dynamic motion and facial expression transfer, as well as consistent 2D cartoon style animation generation without fine-tuning. With MagicDance, we can precisely generate results with consistent appearances, while original T2I models (such as Stable Diffusion and ControlNet) struggle to accurately maintain subject identity information. Furthermore, our proposed module can be viewed as an extension/plugin for existing T2I models without modifying their pre-trained weights.