It is of the significance to organize the knowledge of ancient literatures so as to deeply explore the knowledge elements in these information resources to highlight cultural self-confidence and promote the great rejuvenation of the Chinese nation. With the aids of advanced information processing methods, knowledge computing can extract knowledge elements from a large number of document carriers, efficiently complete knowledge acquisition and knowledge modeling, and achieve the expected goal of knowledge organization of ancient literatures. This paper takes Zipf's law as the theoretical basis of ancient literature knowledge computing, develops a set of word frequency turn point identification standards. Due to the long tail distribution of words in ancient literatures, a frequency position statistics procedure is conducted to deal with words with the same frequency to improve the efficiency of knowledge organization of ancient literatures. This research in clarifies the theoretical basis and specific basic indicators of ancient literature knowledge computing, and provides a reference for the follow-up work of knowledge organization.