The statistical method is used to analyze the spatialtemporal distributions of AMDAR (Aircraft Meteorological Data Relay) data. The AMDAR data are also compared with the conventional radiosonde data. The results show that the AMDAR data is one kind of observation with high spatialtemporal density; the quality of the AMDAR data is close to the radiosonde data; and the temporal density of the AMDAR data is nearly ten times of that of the radiosonde data. Therefore, the AMDAR data can be used to monitor weather systems and the fine characteristics of their evolution around the airports, to provide more usefull references for shorttime forecast, especially for nowcasting.