Abstract: Understanding the interplanetary propagation of solar storms is the foundation of space environmental forecasting and services. The visualization of numerical model results is an important method to analyze the propagation dynamics process and verify the validity of the model. In order to facilitate the visual analysis of the model results, a transfer function design algorithm for volume rendering of time-varying data(TVTF) is proposed. The algorithm is based on the KNN background subtractor method to extract images including motion regions and motion subsets, and then use frequency-tuned (FT) salient region detection algorithm to detect coronal mass ejection(CME) in motion area images, and according to the CME detection results, a color inverse mapping algorithm is designed to find the boundary threshold between the CME and the background. Finally, the transfer function is adaptively adjusted based on the threshold to realize the fast 3D visualization of CME in the motion region at each time step. The experimental results show that the transfer function can adapt to the numerical model results in static and dynamic backgrounds. Compared with the linear transfer function, the occlusion of the line of sight direction is effectively avoided, the change of relative momentum is intuitively and automatically displayed, and the evolution process of CME in interplanetary space is traced. The extraction of local regions reduces data redundancy, and the process of adaptively adjusting the transfer function by automatically analyzing the data with the help of algorithms avoids the inefficiency of manual adjustment.