The fluorescence pharmacokinetic parameters (permeability, etc.) based on dynamic diffuse fluorescence tomography (DFT) can provide reference for studying the dynamic physiological process and pathological information of different biological tissues. The adaptive extended Kalman filter (AEKF) (a method for dynamic analysis) has many advantages, such as precise modeling and multi-parameter online estimation. In this work, the metabolic process of indocyanine green (ICG) in healthy mice liver and tumor-burdened mice subcutaneous transplanted tumor tissue was measured using the dynamic DFT system. Then, we obtained the time-series fluorescence yield images of ICG by DFT method. On this basis, the time-series permeability images of ICG were reconstructed by the AEKF method based on a two-compartment model. The two experimental results verify that the permeability parameters Kpe and Kep of ICG in the tumor are less than that in the liver. Furthermore, the time-series permeability images of ICG also demonstrate that the AEKF method can effectively obtain the real-time and stable ICG pharmacokinetic parameters of complex organisms.