In order to explore the coupling relationship between the industrial economic development and power big data, a vector auto regressive (VAR) model is constructed between the electricity consumption of high-energy-consuming industries and the data of multiple economic indicators. By using the Granger causality test method, industrial economic index data are extracted that has a significant impact on electricity consumption forecasting. Based on this, an Auto Regressive Distributed Lag (ARDL) model of electricity consumption is established in high-energy-consuming industries that takes economic factors into account. An example analysis of industrial electricity consumption and economic data in a certain region from 2016 to 2020 shows that the Granger causality test can effectively dig out the economic indicators related to the electricity consumption of subdivided industries. Considering these economic factors in the regression model, the prediction accuracy of the model is effectively improved.