UniAnimate is a unified video diffusion model framework for character image animation. It maps reference images, pose guidance, and noisy video to a shared feature space, reducing optimization difficulty and ensuring temporal coherence. UniAnimate can handle long sequences, supports random noise input and first-frame conditioning input, significantly improving its ability to generate long-term videos. Additionally, it explores alternative time modeling architectures based on state-space models to replace the original computationally intensive time Transformer. UniAnimate achieves superior synthetic results compared to existing state-of-the-art techniques in both quantitative and qualitative evaluations, and can generate highly consistent one-minute videos through iterative use of the first-frame conditioning strategy.