Holidays，especially long holidays，have a significant impact on the curve of power consumption． The characteristic parameters of industry power consumption fluctuation during the Spring Festival are put forward，and the curve clustering analysis based on K － means algorithm is carried out，which reveals the internal relationship between the daily power consumption of main industries and service industry and the long Spring Festival holiday． The results show that the main industries can be divided into four categories，and the main service industry can be divided into three categories． The power consumption of some industries is basically not affected by the Spring Festival，and among the industries that are obviously affected，there are U － shaped curves and V － shaped curves，some of which have a large decline but rapid recovery，some of which have a small decline but slow recovery，and some of which are highly overlapped with the Spring Festival holiday．