Abstract: In order to solve the problem that the traditional grey prediction method is only suitable for a single variable, a GM（1,n） model with several variables is established on the basis of the traditional GM（1,1） model, and the method of assigning a certain parameter in the traditional model is improved. Re-establish the improved grey forecasting model, and apply the improved GM（1,n） model to power load forecasting.Take the grid load research of a city from 2013 to 2019 as an example, use the GM（1,1） model and the GM（1,n） model to simulate and calculate separately, compare the simulation results of the two, and verify the prediction results obtained by the improved grey model With higher accuracy and smaller error, it is more suitable for multivariate load forecasting.