Abstract: Seed inspection is one of the most urgent problems in the development of China’s seed industry. In this paper, A dynamic quality inspection and grading system of corn kernels was developed, the morphological features were used to divide corn kernels into four grades and the color features were used to grade corn kernels into three levels, the average grading ratio of system were respectively 81.8% and 93.04%; A maize varieties recognition system based on pattern recognition was developed, the system could realize Nongda 80, Nongda 108, Gaoyou115, Nongda 4967 and Bainuo 6 with a average recognition accuracy at 92%; A haploid maize grain grading system was developed, haploid maize seeds were recognized by color features and pattern recognition technology, and a 2 degrees of freedom parallel robot in suction mode was used to sorting with a sorting accuracy of 80%.