Discharge profiles in Lithium-Ion batteries allow the extraction of information about their behavior in the face of different variables such as load, operating time, temperature, among others. One of the ways that initially allows describing the operation of the battery is the output voltage at its terminals. It is possible to expand the capacity of the accumulator and extract more from its operating conditions. This paper presents the modeling of the 18650 lithium cell at different discharge profiles. Initially, the extraction of the current, voltage, and temperature variables from the LG18650HG2 dataset was carried out, which was preprocessed and normalized. Then different deep neural network models were implemented, obtaining a base model and an evaluation model. The first results of the predictions demonstrated better performance of the deep neural network with the Huber function, which only required 50 epochs to obtain losses of less than 2%. With hybrid models, which combine several computational intelligence techniques, the performance of predictions measured as a function of MSE and variance was improved.