In order to research the validity of using the AMDAR (Aircraft Meteorological Data Relay) data in the GRAPES (Global Regional Assimilation and Prediction System) numerical forecast system, the bias and standard deviation of the operational AMDAR data from the National Meteorological Information Center (NMIC) were calculated, and the impact of the AMDAR data on GRAPES was analyzed. We conducted our study using data from May to June 2013 (two months), with an assimilation window of 6 h and a validity period of 7 d. The bias of the temperature was 0.2 ℃, and the standard deviation was 2 ℃. The bias of the wind was 0.2 m/s, and the standard deviation was 3 m/s. The AMDAR data error distribution is almost normal. The results show an improvement in the assimilation and forecast by GRAPES when using the AMDAR data.