The availability of electric energy is one of the clearest indicators of economic development in the world. It is used as a sign of economic growth in times when the world demand for electricity is growing at a dizzying pace. Due to the high price of energy and knowing that the environmental problems are becoming increasingly serious, energy consumption is one of the most critical problems that are being addressed globally under different approaches, both academics and from the industry. The following article presents a basis concept for the definition of a prediction model for consumption and power generation using artificial neural networks. In order to implement the model, the data of the daily energy demand of the country Colombia were used from the first of January of the year 2000 until the 30 of December of the year 2017. The obtained model is characterized by having 13 inputs and a hidden layer with 26 neurons, the training algorithm used is Bayesian Regularization. Finally, the results obtained, as well as the conclusions and future work are presented.