Characteristic Terahertz (THz) region closely related to corn seed aging were screened by using THz time domain spectroscopy technique and moving window correlation coefficient method. There were three different varieties of corn seeds selected and processed by artificial aging for 0d, 1d, 2d, 3d and 4d. The embryos and endosperms of corn seeds samples were isolated and pulverized. THz absorbance spectra (0.2~80cm-1) of the powder samples were collected by THz time domain system and attenuated total reflection (ATR) accessory. The moving window correlation coefficient method was applied to screen the characteristic spectral regions which were sensitive to seed aging. After comparing the parameters of the moving window correlation coefficient, the window width was determined to 10 and threshold value was 0.3. According to the result, the selected characteristic spectral regions showed significant differences among different varieties of samples, which indicated that the seed aging process was greatly affected by the variety. However, there were minor common regions in the same aging stage which were mainly concentrated in 60~80cm-1. It was showed that the similar components were changed at the same time. Due to the selected characteristic spectra of endosperm and embryo at the same aging stage were different, it was showed that the endosperm and embryo had different physiological changes during seed aging process. The experiment result indicated that THz time domain spectroscopy combined with chemometric method can be used to characterize and detect the dynamic changes of corn seed aging rapidly and macroscopically. 