Equatorial Plasma Bubbles (EPBs) are cavity structures with low electron density formed in the low-latitude ionosphere after sunset. Their evolution process can lead to the scintillation and attenuation of radio signals. Precise prediction of the evolution of Equatorial Plasma Bubbles is of great significance in the fields of space weather research and satellite communication. This paper proposes an EPB evolution prediction model based on the SimVP (Simpler yet Better Video Prediction) framework. By learning the spatiotemporal evolution characteristics of EPBs from historical airglow image data, it achieves accurate prediction of future evolution. Through systematic experimental analysis of the influence of key parameters on the model performance, the results show that when the time resolution is set to 3 minutes and the architecture with 6 input frames and 6 output frames is adopted, the model performs optimally (SSIM = 0.989, PNSR = 34.704). The complexity of the spatial morphology of EPBs has a significant impact on the prediction accuracy, while the interference of light pollution is relatively limited. This model not only provides a data-driven and efficient prediction tool for the evolution of EPBs, but also offers technical support for the restoration of contaminated airglow observation data.