Abstract: The concept of data fusion was proposed in 1970s and developed in 1990s. Many researches have been done in this area. However, for a long time there is not a specific definition of data fusion that is commonly accepted and used by academic societies of information technology. The description of data fusion is usually different from one to another, which confuses the communication in the scientific community and influences the technology transfer to industry. On the other hand, how to apply the data fusion technology achievements to remote sensing in agricultural condition monitoring is also a very important area. In this paper, the concept of data fusion and its development are reviewed in detail, including the definition, fusion levels, and methods. Then the paper analyses the status of its application in agricultural condition monitoring: high spatial and spectral resolution images begin to be used in agricultural condition monitoring together with low resolution ones, and some ancillary data play a more important role in helping with crop identification and image classification. For how to better use data fusion in agricultural condition monitoring, which is not only operational, on large scale, with great complexity, but also requires high accuracy, the application prospects are discussed by the authors.