Using the statistical analysis, 13 factors that affect the construction’s carbon emission intensity is summarized based on relevant data of three regions including Beijing, Tianjin, and Hebei were collected. The stepwise regression analysis was used to find the 6 key factors with the 5-Fold Cross-Validation to solve the problem of data shortage. the artificial neural network was used to predict construction’s carbon emission intensity in Beijing-Tianjin-Hebei. The sensitivity analysis is carried out to make the model more understandable and then found the 4 most important factors. The results show that an accurate prediction has an accuracy of about 99%. According to the key factors and the most factors, propose suggestions for energy conservation and emission reduction in the construction industry.