At present，due to the impact of economic development and environmental constraints，the “de-capacity” and “environmental protection” actions of traditional high-energy-consuming industries such as steel and non-ferrous metals have been intensified. Due to the lack of effective monitoring techniques，In practice，there is often a phenomenon that is not in place. This paper uses the power load big data of key industrial users such as steel，through the combination of wavelet packet decomposition and fast Fourier decomposition，to construct a non-intrusive monitoring method to achieve key production process monitoring for high energy-consuming and high-pollution industrial enterprises. Taking a steel plant in Henan as an example，using the constructed non-intrusive monitoring method，the load curve of the main production process is decomposed to achieve effective monitoring of the main processes of iron making，steel making and steel rolling. At the same time，through the typical link power capacity measurement，the company’s steel production capacity is obtained，providing technical support for the supervision of key industrial users such as steel.