China is stepping into a new normal economic development， with the economic structure and the energy system adjusted extensively. As a signpost of economic development， the power demand growth has also changed with the uncertain environment. In such a complex situation， it is of great significance to guarantee accurate early warning in the medium and long term power demand for the integrated resources planning of power system. To this end， the theoretical framework of medium and long term power demand warning was proposed in energy internet. Firstly， the theoretical framework of medium and long term power demand warning was constructed. Then， the medium and long term power demand warning methods were proposed based on the internal analysis and external factors based analysis， and several early warning models such as correlation analysis， prediction analysis， and classification analysis were introduced. Finally， the future researches on medium and long term power demand warning were summarized and prospected.

early warning, big data, warning probability, power demand, energy internet