Abstract:In view of the non-linear and non-stationary characteristics of the ionospheric total electron content(TEC) data, based on the auto-regressive moving average(ARMA) model prediction method and the empirical wavelet transform(EWT) method, we propose a combined short-term ionospheric prediction method. Using the ionospheric TEC grid data provided by IGS, the results show that, compared with the average relative accuracy of single ARMA model, the average relative accuracy of the combined model in low solar activity and high solar activity years is increased by 4.8% and 2.8% respectively, and the average relative accuracy of the combined model in the first day is increased by 7% and 6.1%.