This paper proposes an artificial neural network (ANN)-based prediction model to accurately estimate green hydrogen production in the Dakhla Oued Ed-Dahab region, using detailed data ex-extracted from HOMER Pro software. This model aims to optimize the management of renewable energy resources in a smart microgrid, particularly with regard to hydrogen production and storage, a key solution for sustainable development. The approach integrates crucial parameters from renewable sources such as solar and wind power, as well as operational data from the microgrid. The results obtained show a high degree of model robustness with reliable predictions. This model can be directly integrated into an energy management system (EMS) to improve decisions related to hydrogen production and storage, contributing to better management of energy surpluses and reduced dependence on fossil fuels. Prospects include improving the model by integrating additional data and considering other environmental parameters.