Abstract: As the installed capacity and generation of renewable energy continues to increase, accurately predicting the green electricity demand of the future electricity market is crucial for planning new energy installed capacity and green certificate trading decisions. Aiming at the problem of limited historical data related to green electricity, this paper proposes a data integrity index, which is combined with convolutional neural network and long short-term memory network to build a new green electricity demand prediction model, which improves the model′s ability to recognize and learn important features in time series data. By analyzing the actual data of Jiangsu Province from 2011 to 2022 and the data expanded according to the data integrity, the superiority of the model in forecasting accuracy compared with the traditional method is proved.