The design of dictionary index structure is the base of Chinese information processing and its properties will influence the effect of Chinese word segmentation greatly. In this paper, a Hash table is firstly established for all commonly-used characters, so that all characters in each word could be found quickly; then for each character, the number of word and their composition relationship in the word consisting of the character are recorded in the word chain to form the all-character index structure; next, the paper discusses the construction and maintenance methods of the dictionary and presents the dictionary constructing, adding and deleting algorithms. Finally a Chinese word segmentation algorithm based on all-character index dictionary is proposed and some comparisons with traditional dictionary in dictionary construction, query speed and function are made.