Abstract: Adopting a mixture variation algorithm for reconstructing the image can not only maintain the continuity of the reconstructed results, but also keep the discontinuity during the process. A two-dimensional numerical model on EIT was established based on the finite element method. To optimize the model, we discretized the numerical model, defined the physical characteristics of the model and discrete objects, imposed the boundary condition, calculated the boundary voltage, simulated the forward problems and determined the objective function of the algorithm. Finally, we carried out the image reconstruction and evaluated the performance of the mixture variation algorithm. The structural similarity of the image were improved by about 2% and 2.9%, and the reconstruction speed were increased by about 2.4% and 2.6% respectively, after reconstruction by the mixture variation algorithm under the appropriate weight value compared with that after reconstructed by the Tikhonov algorithm and the total variation algorithm. The mixture variation algorithm can be both used in the image reconstruction of clinical measured data and in the simulation researches and experiments.