In order to achieve higher recognition rate of unfertilized eggs, the paper proposes a new method based on the area feature of egg's vertical image and the classifier BP network. The eggs hatched 2-5 days are selected as the experiment objects in our study. Four kinds of feature, including the gray feature and the area feature extracted from the egg's vertical image and the egg's horizontal image, respectively, are sent to BP network to accomplish the recognition of unfertilized eggs. The experiment results indicate that the area feature of egg's vertical image contains more category information, and this kind of feature classified by BP network can reach 97.86% of recognition rate. Moreover, it is almost free from some external factors, such as light, camera, eggshell and the parameters of BP network. While the other features sent to BP network, the recognition rate can be reached to only 86.43%. This study provides an effective and useful method to recognize unfertilized egg in early incubation stage.