Wire arc additive manufacturing (WAAM) is gaining increasing attention thanks to its advantageous features, such as high deposition rate, high material usage efficiency, and low costs of initial setups. In this process, the single weld bead and melting pool size play a crucial role for the printing stability and simulations. In the current research, the prediction models of weld bead dimensions and melting pool size that can be integrated into numerical simulations in the WAAM process of SS 316L will be developed. For this aim, a series of experiments was performed based on Taguchi-L16-orthogonal-array approach with the three input variables (i.e., welding current I, voltage U, and welding speed v). After the experimental runs, the data on the weld bead and melting pool sizes was collected to develop ANN (artificial-neural network) models. The results show that the R value of all the models is 0.99, indicating that all the ANN-based models of bead width, bead height and melting pool size have high accuracy and they can be used to predict the output attributes with high confidence. These models can also be used for path planning and optimal process parameter estimation.