Along with the construction of smart distribution network, power grid data resources begin to grow rapidly. Tremendous amount of data is scattered in different systems and irregular in format. Because of the low utilization rate of data information, distribution network planning still relies on artificial methods and subjective judgment. This paper proposes a data fusion planning framework gathering "power flow", "business flow" and "information flow" together. Distribution network status diagnosis, weak links identification, load forecasting and project design assistant are realized through modular development. Afterwards, the planning project database can be generated. This planning framework helps to improve the accuracy and rationality of the planning. Its application in the distribution network in State Grid Shaanxi Electric Power Company will be shown in the end of the paper.