This paper explores a new method to model and forecast the global earthquake energy release time series. The ISC-GEM catalogue of global events with magnitude Mw≥6.4Mw≥6.4 is used in this study. The magnitudes of individual events are converted into seismic energy using an empirical relation. The annual earthquake energy time series is constructed by adding the energy releases of all the events in a particular year. Then, the energy time series is decomposed into finite number of intrinsic mode functions (IMFs) using empirical mode decomposition (EMD) technique. The periodicities of these IMF’s and their contribution to the total variance of the data are examined to identify the influence of natural phenomenon on earthquake energy release. The artificial neural network technique (ANN) is further used for modeling the energy-time series. The model is verified with an independent subset of data and validated using statistical parameters. The forecast of the annual earthquake energy release is provided for the year 2016.