A near real-time Global Land Data Assimilation System (GLDAS) has been successfully implemented in the NCEP recently. Its purpose is to ingest satellite- and ground-based observational data products, using advanced land surface modeling techniques to provide the accurate fields of land surface states for weather and climate predictions. Previous studies show that the NCEP Global Forecast System (GFS) tends to overestimate the precipitation over eastern half of the continental United States (CONUS) in warm season, which results in cool surface temperature. The CPC Merged Analysis of Precipitation (CMAP) is used as precipitation forcing to produce the land states, such as soil moisture, in GLDAS. With CMAP precipitation the more accurate land states are expected compared to the ones provided by the data analysis system in the NCEP GFS which use the model precipitation. In this study 31 6-day forecasts of the NCEP GFS for the summer period were carried out by using GLDAS land states as the initial conditions. The results shows modestly positive impacts on precipitation skill score and 2-m air temperature bias; and virtually neutral impact on 500 mb height anomaly score.