In this study, We construct the EMD-LSTM model, combined the Empirical Mode Decomposition algorithm (EMD) and the Long Short Term Memory neural network (LSTM), to predict the variation of the >2MeV electron fluxes. The Pc5 power and related geomagnetic indexes as input parameters are used to predict the >2MeV electron fluxes. Compared the prediction results of the model with other classical prediction models, the results shows that the one-day ahead prediction efficiency of the > 2MeV electron fluxes is above 0.80, and the highest prediction efficiency can reach 0.92 in 2011-2013, which is much better than the prediction result of classical prediction models. Selected two high-energy electron flux storm events to verify, the results indicates that the performance of the EMD-LSTM model in the period of the high-energy electron flux storm is also relatively good, especially for the prediction of high-energy electron fluxes at extreme points, and the prediction is closer to actual observation.