Grid planning is a new planning model proposed in the context of the increasing scale of distribution networks. According to the principle of grid division, the power supply area is divided into multiple power supply grids, and each grid is separately planned. Considering the characteristics of grid planning and the limitations of traditional reliability assessment methods, a prediction model based on reliability historical data is proposed, which is mainly used to predict the reliability of the grid framework with relatively insufficient original data information. Firstly, the key factors affecting the power supply reliability are selected by the grey correlation analysis method as the input of the prediction model. Then, in order to improve the accuracy of the prediction model, three scientific commonly used prediction models, including regression prediction method, grey prediction method and artificial neural network method are used to predict the reliability index respectively. Finally, an example analysis of a grid planning area shows that the proposed model can effectively predict the reliability of the planning grid in a certain period of time.