Abstract: The in-depth portrait analysis of the power consumption characteristics of commercial power users purchasing electricians will provide theoretical basis for power forecasting, power generation planning and policy system improvement. This paper obtains the annual, monthly and daily electricity consumption data of commercial electrician customers in different industries and regions, and synchronously obtains the matching meteorological, electricity price, economy, holiday arrangement and other data. Through the data processing algorithm, the outlier points are eliminated from the sample data, the missing values are filled out, and the correlation between electricity consumption and temperature, economy, electricity price and holidays is established, and the feature vector matrix of commercial customers who purchase electrical agents in full volume is constructed. The feature vector matrix is analyzed by clustering algorithm, and the user characteristics of key industries are sketched with the completed clustering. The sketching results can provide reference for practical production.