Abstract Aimed to be ready for the transition from research to operation, we have developed a solar wind model by coupling a data-driven empirical coronal model with a magnetohydrodynamics heliospheric model. We performed a data-driven simulation of the solar wind for a two-year period during the declining and minimum phases of solar cycle 23. Comparisons with OMNI and Ulysses spacecraft data show that the model can reproduce the large-scale variations of the solar wind plasma parameters. The evolution of geocentric solar magnetospheric (GSM) B x and B z components are also reasonably duplicated by the model in terms of polarity and strength. Apparent signatures of the Russell–McPherron (R-M) effect are found from both observed data and simulated results, indicating that during the investigated interval the R-M effect is the dominant mechanism that controls the large-scale evolution of the north–south component of the interplanetary magnetic field in the GSM frame. The results demonstrate that the established model can provide valuable space weather information about the solar wind.