In this article, a predictive model for ground motion characteristics is developed using the artificial neural network (ANN) technique. This model is developed to predict peak ground acceleration (PGA), peak ground velocity (PGV), peak ground displacement (PGD), spectral acceleration at 0.2 and 1 s. The input parameters of the model are moment magnitude (M_w), closest distance to rupture plane (R_cd), shear wave velocity in the region (V_s30), and focal mechanism (F). The updated NGA-West2 database released by Pacific Engineering Research Center (PEER) is employed to develop the model. A total of 13,678 ground motion records are used to develop the model. The ANN architecture considered in the study has four input nodes in the input layer, three neurons in the hidden layer, and three output nodes in the output layer. The ANN is trained by a hybrid technique combining genetic algorithm and Levenberg–Marquardt technique. The results of the study are found to be comparable with the existing relation in the global database. The model developed can be further used to estimate seismic hazard.