Smart agriculture is quite common in all parts of the world. It is necessary in this modern world to meet the basic human needs. The latest technologies play a major role in different stages in agriculture process. In the classical method of agriculture human intervention is expected in every stage to get the desired yield. It is proposed to develop a system, so that the land owners and farmers observe the necessary parameters from the farm land with the help of latest technologies and take appropriate actions. To develop such system, virtualizing each stage of a plant is required. Through this virtualisation plant parameters can be monitored. This technology is called Digital Twin (DT). The DT technology has wide applications in various domains and it develops the virtual model of any physical system. It’s an emerging area in the science and engineering domain. This technology is proposed to be implemented in the agricultural sector to aid farmers and landowners. It supports various analysis during the different stages of plant growth. In future DT will play major role in smart agriculture. DT technology is applicable in various stages in agriculture. This paper explores a particular aspect of the DT, which detects the diseases of a plant through its leaf. Lots of real-world data are required for diseases detection model development. Public repositories are used in this research to create an image data store and to develop deep learning classification model. The developed model helps the society to bring revolutions in smart agriculture.