Abstract: This paper selects the typical industry data of copper and copper consumption in China in the years 1999-2015and applies the adaptive-Lasso method to analyze and identify the key industries that affect the consumption of copper in China.Based on the analysis,a combined model of grey Verhulst and BP neural net-work is constructed,and the consumption of copper in China is predicted by this model.The peak of copper consumption in China will arrive in the years 2020-2026,with a peak range of 990×10⁴-1300×10⁴ tons.And we suggest that the copper smelting production capacity should be controlled within 1280×10⁴-1690× 10⁴tons,and the new smelting investment should be suppressed so as to avoid overcapacity. 

Key words: Copper consumption peak, Adaptive-Lasso, Grey, Verhulst, BP neural network