Pest identification plays an increasingly important role in agricultural production, and effective control measures based on the results can reduce the impact of pests and diseases and ensure higher and more stable yields. Advances in machine learning and deep learning, especially in image recognition, provide a better way to identify pests and diseases. EfficientNet-B3 neural network was used to identify and classify healthy crops and diseased plants with 98% accuracy, and data enhancement technology was used to address the insufficient amount of data.