To improve the efficiency and accuracy of corn disease identification, and address the shortcomings of Convolutional Neural Network (CNN) models—such as low recognition accuracy, weak generalization ability, and poor robustness—when the disease sample dataset is insufficient, a new corn disease identification model is proposed based on the Inception V3 network model and transfer learning. The Inception V3 model was pre-trained on the corn disease dataset of the ImageNet library without freezing all layers. Subsequently, its fully connected layers were removed, and a new fully connected layer structure was redesigned to form the corn disease identification model. The self-built dataset was used to train, validate, and test the proposed model, and the influence of initial learning rate, data augmentation, and transfer learning methods on the model’s performance was analyzed. The experimental results show that the proposed model achieves a recognition accuracy of up to 97.76% for 6 types of diseases. Compared with the comparison models, it has the advantages of fast convergence speed, strong generalization ability, and high accuracy, along with good robustness.