The increasing use of distributed generation in the Electric Power Systems has brought these several problems, since the controllers of these systems must take into account one more variable for any decisions. Within this context, there was a special emphasis on the use of electric energy storage systems to reduce losses, with the allocation of them as a load at the overgeneration points and to reduce system losses at peak demand points. This article presents an analysis of the distributed generation to compare the electrical losses using electrical energy storage. The technique used to estimate the electric energy generation of a photovoltaic system in one day, according to temperature and irradiation data, was the application of artificial neural networks. Thus, the storage element appropriate to the demand and generation levels was first defined and the OpenDSS tool was used to perform the study of the load flow and to measure the effect of the use of these storage systems in the IEEE 37 bus network. Using energy storage in this application has proved to be advantageous, reducing the losses of active power at the peaks and reducing over-generation during the day, measured as crucial for the future of the distribution.