Carbon emissions from the power sector account for about one-third of China’s total carbon emissions, and carbon emission reduction in the power sector is crucial to the realization of the “double carbon” goal. This paper proposes a prediction model for grid carbon emission factor based on the combination of multiple linear regression models and the GM(1,1) model. The grid carbon measurement model is built using the theory of carbon emission flow, and the grid carbon emission accounting model is built using the consumption side. The average grid carbon emission factor is calculated by dividing the grid carbon emission factor into three different dimensions. The multiple linear regression model is used to study the correlation between the independent variables and the dependent variables, and the specific values of each regression factor are predicted by combining with the GM(1,1) model, and it is judged whether the prediction model is reasonable or not. Among the contributions of the increase in carbon emissions from the power grid, the contribution of electricity consumption increased from 2010 to 2014 and reached a peak of 2.9824 million tons in 2014, and the carbon emission factor value of the power grid gradually decreased from 0.719kg CO2/(kW·h) in 2010 to 0.593kg CO2/(kW·h) in 2022. The MR-GM(1,1) model is applied to the prediction of carbon emissions from power grids, and the absolute error of the results is within 15,000 tons, and the maximum relative error is only 2.42%. The calculation and prediction of carbon emission factors of the power grid can help power grid enterprises to clarify the trend of carbon emission, which is conducive to the realization of the low-carbon goal of “carbon neutral” and “carbon peak”.