The automatic recognition algorithm of emergence and three leaf stage of maize is proposed in order to dynamically monitor the growth of maize in realtime and provide accurate information about growth status for farming activities. This algorithm realizes segmentation of corn images and extract image characteristics such as contour and skeleton. To determine whether the corn has been in emergence and threeleaf stage according to the extracted image features. This algorithm and computer vision class library openCV are used for the design of emergence and threeleaf stage detection system and the target of detecting emergence and threeleaf stage of maize is achieved automatically. What’s more, in the environment of VS2013, it achieved the development of this detection system. The speed of this detection system is fast and the test results are accurate. It can be used as the development foundation of all maize growth period detection system.