Since the daily load power curves only considers the instantaneous power characteristics of the data point, the larger consumers' load disturbance will affect the ability of the daily load curves to describe the actual load characteristics. Taking the daily load power curves as the only basis to represent the actual load characteristics of consumers may cause misjudgment and bring errors to the description of typical industry load characteristics. In order to solve such problems, this paper uses the traditional K-means clustering technology to obtain daily electricity consumption curves based on the power user electric energy data acquire system, and then obtains more accurate typical industry cluster center curves, and proposes a classification and synthesis method of typical industry load characteristics based on consumers' daily electricity consumption curves. The results show that compared with the traditional clustering method of daily load power curves, the proposed method has certain advantages in clustering results, clustering quality and clustering accuracy.It overcomes the limitations of daily load power curves itself, has the advantage of reflecting the overall trend of load change, and can truly reflect the electricity consumption characteristics of consumers in this area, so it has a good application prospect.