Abstarct: Enterovirus 71 is one of the major human pathogenic viruses that can lead to severe neurological and respiratory complications, and no widely available specific and effective antiviral treatment has yet been established. In the present study, with the aim of identifying and analysing the structural features influencing antiviral activity, a series of pyridyl–imidazolidinone derivatives were investigated as potential inhibitors of Enterovirus 71. The molecular structures of the compounds were first designed in three dimensions and optimised using quantum chemical methods. Subsequently, various physicochemical, electronic, and structural molecular descxriptors were extracted and, after preprocessing and removal of correlated descxriptors, were employed for quantitative structure–activity relationship (QSAR) modelling. QSAR models were developed using multiple linear regression and feedforward artificial neural networks with the backpropagation learning algorithm (MLR–ANN). Model performance and validity were evaluated using statistical parameters including mean squared error (MSE), mean absolute error (MAE), coefficient of determination (R²), leave-one-out cross-validation coefficient (Q²LOO), Y-randomization test, and residual plot analysis. The results demonstrated that the developed models possess satisfactory capability in describing and predicting the antiviral activity of the compounds, and that several electronic and spatial descxriptors play a key role in viral inhibition. Furthermore, molecular docking studies were performed to investigate the interaction of selected compounds with the molecular target. Overall, this study provides an efficient frxamework for the design and optimisation of novel Enterovirus 71 inhibitors and may serve as a basis for the development of effective antiviral agents.