<p>In this study, the stochastic finite-fault method based on dynamic corner frequency is employed to simulate Tangshan Guye M_S5.1earthquake. Due to limited seismic data, establishing a set of input parameter models consistent with the earthquake’s conditions is essential for improving simulation accuracy. Four models were developed by varying three input parameters: stress drop, kappa, and dip angle. The simulated acceleration time series and acceleration response spectra from each model were compared with actual observations to identify the most accurate parameter combination. Further validation was conducted by comparing the selected model with two additional models. Results reveal discrepancies in duration and waveform between simulated and observed acceleration time series across different parameter sets. Stress drop and kappa values significantly influence simulation outcomes: an increase in stress drop elevates peak ground acceleration (PGA), while higher kappa reduces PGA. The dip angle exhibits minimal impact. Model 3, which demonstrates the closest alignment with observed data, is selected as the optimal reference for ground motion simulations of the Tangshan Guye earthquake. The acceleration time series simulation results of Model 3 differ from the actual observed data by only about 1 gal. Compared to the other models, the model bias of Model 3 is less than 0.2, and its variation curve is smoother. The synthetic ground motions generated in this study provide a foundation for post-disaster seismic design in the Tangshan region. They can be used to predict future earthquakes in the region, helping to reduce the severity of earthquake-related disasters.</p>