For Cornell model in the same scintillation conditions, the longer the data update period, the greater the deviation of the estimated and model input amplitude scintillation index S4, and with the enhancement of ionospheric scintillation, the ionospheric scintillation model based on Cornell model with adaptive S4 was proposed, namely AS4-Cornell model. The model used the deviation of S4 as the feedback of the back propagation (BP) neural network proportional integral derivative (PID) algorithm to automatically adjust the zone weights of the input complex Gaussian white noise so that the final generated scintillation signal met the index of the scintillation index S4 input to the model. The results show that the amplitude and phase scintillation sequence probability distributions of the AS4-Cornell model are consistent with the ionospheric scintillation theory during the simulation, and the maximum deviation of the estimated ionospheric amplitude index S4 from the model input S4 index is 0.001. The maximum deviation of the ionospheric amplitude index S4 estimated by the AS4-Cornell model from the model input S4 index for the global positioning system (GPS) ionospheric scintillation simulator testis 0.09. Compared with the Cornell model, the ionospheric scintillation signal generated by the AS4-Cornell model can better reflect the intensity of the ionospheric scintillation index S4 inputted by the model.