Angiogram video sequence compression based on the wavelet representation and fast motion estimation is presented in this paper. The characteristic motion of angiograms is investigated, leading to a statistical result that is analyzed for fast motion estimation in order to get good interframe prediction. Wavelet transform decomposes the video frame into a set of sub-frames in different resolutions corresponding to different frequency scales. Weighted scale and uniform scalar quantization are applied to the subband coefficients according to their perceptual importance to the reconstructed image. Arithmetic coding is also used to achieve high coding efficiency for either base frame or error frame. The proposed algorithm not only achieves low bit-rate coding but also maintains high fidelity which is essential for medical images.