By introducing the concepts of characteristic rate curves and rate curve decomposition, a novel framework for rate-distortion (R-D) analysis and source modeling is developed in this work. With this framework, a fast algorithm is then proposed to accurately estimate the R- D curve of wavelet-based image coders. The proposed algorithm is applied to the SPIHT (Set Partitioning In Hierarchical Trees)1 and Stact-Run (SR) encoders2. Our extensive experimental results show that the relative R-D curve estimation error is less than 5%.