Abstract: The present invention provides a global ionospheric total electron content prediction system based on a spatio-temporal sequence hybrid framework. The prediction system implements computational processing for two types of spatio-temporal sequences, wherein for a stationary spatio-temporal sequence, a STARMA model prediction method is constructed in the present invention; for a non-stationary spatio-temporal sequence, a nonlinear spatio-temporal trend is firstly extracted from the non-stationary spatio-temporal sequence by adopting a ConvLSTM method until the extracted residual passes a stationarity test, and then the electron content is predicted using the STARMA model prediction method.