The invention takes a KNN algorithm as the core to realize temporal classification of mass inventory materials in a material storage process of an electric power company according to a power grid material in-stock situation and through dimensions of prefectural and municipal unit, material type, movement type, factory type, warehousing batch number, storage time, unit price, aggregate amount, current date and material warehousing data. Through the classification method provided by the invention, characteristic data classification of electric power material in-stock inventory is realized, material storage time and regions are divided, and normality, expiration and time-limit-exceeding identification of inventory time is effectively achieved.