Seed maturity is an important factor affecting seed quality and crop yield. Therefore, rapid nondestructive testing of single seed maturity is an important basic technical guarantee for promoting modern single seed sowing technology for improved seeds. A rapid identification model of single corn seed maturity was established by combining nearinfrared spectroscopy and chemometrics. The experiment collected the near infrared spectrum of 200 single corn seed samples. Interference information was reduced by particle morphology, stray light, etc. during single seed collection. The segmentation spectral region was used respectively, and the continuous projection algorithm was used to screen the feature wavelength optimization to establish a seed maturity determination model. The experiments showed that the recognition accuracy of the SVM based single corn seed maturity discrimination model established in the 5500~4000cm-1 spectral region could reach 92% after the Savitzky-Golay 5point convolution first derivative pretreatment. The results showed that NIR had a bright application prospect in the rapid and nondestructive discrimination of single corn seed maturity.