In order to meet the challenge of Big Data, enhance the intelligent level of distribution grid, the better for the power user service. Starting from the characteristics of 4V of Big Data, the 6 links to the power supply system of industrial chain (i.e., planning, design, construction, operation, management & regulation of distribution grids, and equipment design and manufacturing) is relatively mature degree angle; the needs of Big Data's application in distribution network have been analysed. By using the method of SWOT, analysed the double-edged sword effect Big Data for the distribution gird, provides both opportunities and challenges. The benefits and opportunities is that, Big Data bringing data view, changing thinking methods and tools, expanding the application scene, providing better service to the society, enhancing the value of the opportunity. At the same time, Big Data will lead to the challenges in distribution grid, for example, because of security challenges of Big Data itself, Big Data more concentrated, cause safety challenges in distribution grid is more serious; the energy consumption challenges of Big Data; Big Data privacy threat distribution grid and user. The demand for Big Data's application in distribution grids in industrial chain is that, from strong to weak, management & regulation of distribution grids, operation, equipment design and manufacturing, construction, design, planning. Big Data's source of power supply enterprise's internal operation, including 3 parts, physical grid operation, marketing services and grid enterprise operation. Power system technology innovation by three wheel drive (experimental science, theoretical science, computational science) increased to four wheel drive (experimental science, theoretical science, computational science, data intensive science/data exploration science) paradigm. Big data is still "explosion", control the fourth paradigm -- data intensive science also need to redouble our efforts. Non structure data in distribution grid will be rapid growth, 50% more than the amount of data in the next five years.