This paper presents the development of a robust ground motion model (GMM) employing machine learning techniques. The objective is to predict the parameters of an improved site-based stochastic ground motion simulation model based on a set of seismological parameters, thereby enabling a scenario-based simulation of ground motion for specified earthquake and site characteristics [1]. The proposed GMM leverages the advantages of machine learning methodologies to establish a predictive framework that enhances the accuracy and efficiency of ground motion predictions [2-4]. The study utilizes a dataset of ground motions with known input seismological parameters, including moment magnitude (Mw), rupture distance (Rrup), average shear wave velocity at the upper 30-meter soil (Vs30), and fault mechanism. Residual analysis is conducted by utilizing a likelihood function derived from the adapted analytical solution presented in [4]. To guarantee the prediction accuracy of the models also for unseen future data, only 80 percent of the data is used for training, and the rest is reserved for testing the trained model. The model hyperparameters are tuned to control bias and variance trade-offs by k-fold cross-validation. Through the application of advanced machine learning algorithms, including artificial neural network (ANN) and random forest (RF), the developed GMM demonstrates its efficacy in capturing complex relationships between input parameters and ground motion outcomes.