Residential building energy consumption has gained attention in terms of energy utilization and efficiency. It is more difficult to obtain detailed energy consumption data of a residential building as compared to an industrial building. Therefore, there has been an increased focus on modeling and simulation for this sector. Studies have shown that neural network (NN) method can be applied to the prediction of energy consumption in residential homes. A neural network model typically consists of three layers: an input layer, a hidden layer, and an output layer. Each layer may contain any number of neurons. The number of neurons in the input and output layers are constrained by the number of inputs and outputs of the model. Therefore, the accuracy of a NN model depends on the number of neurons within the hidden layer. In the development of a NN network model using energy and weather data for a research house at The University of Texas at Tyler, the number of neurons in the hidden layer had to be analyzed to optimize model performance. A suggested equation from literature based on empirical analysis evaluates the optimal number of neurons in the hidden. Given 3 inputs, the optimum number of neurons in the single hidden layer was found to be 7 neurons. This number was verified by comparing the models of different number of neurons where the optimal neuron model has the maximum coefficient of determination of 0.878.