This paper introduces an innovative methodology for predicting the maximum dynamic response of structures using capacity curves and artificial neural networks (ANNs). This novel approach offers a quick and accurate procedure for estimating target displacements, obviating the need for intricate supplementary computations. The method generates a comprehensive dataset encompassing the bilinear representation of a single-degree-of-freedom (SDOF) characteristic, with ground motion parameters as inputs and maximum inelastic displacement as the corresponding output. This dataset is used to train an ANN model, with meticulous calibration of hyperparameters to ensure optimal model performance and predictive precision. The findings of this study demonstrate that the ANN model showed operational efficacy in approximating dynamic displacements. It is notably revealed that the size of the dataset significantly influences the ANN's performance and predictive accuracy. Through comparative analysis with established methodologies such as the displacement coefficient method and the modified coefficient method adopted by the Federal Emergency Management Agency (FEMA), the ANN model emerges as a fast tool for precisely predicting the dynamic response of single-degree-of-freedom systems, particularly those characterized by vibration periods exceeding 0.5 seconds. Consequently, this research culminates in the assertion that the ANN, owing to its inherent simplicity and impressive precision, is an alternative tool for estimating target displacements.