Abstract:The electricity consumption as a key indicator for objectively measuring the warmth or coolness of the economy plays a crucial role in formulating energy and economic policies. The changes in electricity demand from the primary industry, secondary industry, tertiary industry, and urban and rural residents have a guiding role in understanding the changes in industrial structure and the future trends of the economy. This paper uses the partial least squares method, combined with quarterly trend decomposition tools and industrial decomposition of electricity consumption, to construct a monthly multi-step forecasting model for China’s overall social electricity consumption. The forecasting results show that the partial least squares prediction model with industrial decomposition of electricity consumption and seasonal adjustment has higher accuracy compared to other commonly used models.