Abstract:A new random walk process considering between-epoch variations of ionospheric delays is employed. GPS data sets collected from 170 globally distributed stations of the International GPS Service(IGS) network in a one-month period(July 2016) are exclusively processed in static and simulated kinematic modes. The convergence time and positioning accuracy are used as indicators to validate the stochastic modeling for ionospheric parameters. For the convergence performance, the proposed model is not affected by the spectral density of the random walk process, while the convergence performance is much better than the traditional random walk process at the small spectral density. Positioning accuracy derived from the proposed model and the white noise process are comparable, with an average RMS of 5 cm in static mode, and 8 cm in kinematic mode.