Abstract: Deep learning is one of the most important technologies in field of artificial intelligence．Deep learning has set off a research climax in academic field and engineering application．In view of application potential and importance of deep learning in agricultural field, concept of deep learning was described in detail through study of deep learning literature, and its characteristics, advantages and disadvantages, variants and application status were reviewed based on structural characteristics of typical deep neural networks．Development and application of deep learning in speech recognition, agricultural scene object detection and semantic segmentation in agricultural field were introduced．Current problems and future research directions of deep learning in agriculture were analyzed．