Abstract: Using the precipitable water vapor (GPS-PWV) inversed from ground-based GPS that based on the advanced ZHD model and localized Ts model, as well as integration hourly meteorological data from automatic weather station, the atmospheric water vapor variation characteristics and evolution features of GPS-PWV during 14 heavy rainfall events are analyzed at Huaihua in 2017. As the results shown, GPS-PWV could reveal the atmospheric water vapor variation characteristics of Huaihua region. The monthly change of PWV-P data pair is evident. The PWV appears a lower value with a smaller range accompanied by a 14.75hPa higher surface air pressure than that in summer when precipitation occurs during winter, which gradually increases with a lower surface air pressure while precipitation during spring. In summer, the PWV rises to the annual peak value with the lowest surface air pressure under rainfall, and it scatters to low value area in autumn. In 14 heavy rainfall events at Huaihua during flood season of 2017, all of the PWV values exceed corresponding monthly mean, besides there is a well corresponded relationship between the maximum rainfall and maximum PWV in hourly scale. Before the heavy rainfall occurred, the PWV increases comparatively distinctly with a clear decrease of the surface air pressure, and that could be a preferably reference point in the local strong precipitation nowcasting.