Artificial intelligence technology has shown incomparable advantages in solving traditional agricultural problems. At present, the determination of moisture content and grade of high moisture corn in grain purchase is still mainly based on artificial sensory detection. There are many problems such as heavy workload, low efficiency, poor repeatability, and strong subjectivity, which will cause the loss of grain quantity and quality, and this problems even affect the benefits of enterprises and farmers. Aiming at the technical difficulties in rapid detection of high moisture maize, this study obtains enlightenment from the experience of front line personnel in visual measurement of moisture. Based on mining the rich information in corn kernel images, a new bionic image rapid nondestructive measurement method using the mutual information entropy as the coupling degree evaluation index, were studied then a intelligent purchase grading system of high moisture corn were developed and tested. The test results show that on the basis of realization of the bionic intelligent algorithm of ‘watching moisture by machine’, the purchase of high moisture corn is a technical path worth exploring.