In this study, we use the MM5 model linked to the OSU land surface scheme to assess the strength of soil moisture-rainfall feedbacks in the NAMS region that result from realistic soil moisture forcing. Simulations are driven by NCEP reanalysis. The horizontal resolution of the finest grid is 30 km. Three member ensemble experiments begin on June 1 and end on October 1. First, we use the coupled MM5/OSU model to simulate the NAMS climate and soil moisture in the wet (1999) and dry (2000) monsoon seasons. Second, we repeat these two experiments but constrain the precipitation rate in July over the entire NAM region so that it approximates the mean state. All of the experiments preserve the temporal variability of the sea surface temperature (SST) in the surrounding oceans. We compare the atmospheric and land surface states in the control and sensitivity experiments. The results show that with the precipitation increasing in July, a positive feedback of soil moisture-precipitation can be found simultaneously over the NAMS region. With time passing after the artificially imposed precipitation anomaly stop, the positive signal in responding soil moisture moves eastward from NAMS region and an opposite pattern appears in September.