The invention relates to a constraint conditional random field-based Vietnamese noun chunk identification method, and belongs to the technical field of natural language processing. The method comprises the steps of firstly constructing a Vietnamese noun chunk corpus library; performing statistics on part-of-speech characteristics of noun chunks in the corpus library, and establishing a constraintcondition; training noun chunk corpora by applying a conditional random field to obtain an initial conditional random field identification model; adding the established constraint condition to obtaina final constraint conditional random field identification model; and according to a noun chunk identification model parameter sequence, identifying Vietnamese noun chunks to obtain a final identification result sequence. The Vietnamese noun chunks are effectively identified; and powerful support is provided for work of lexical analysis, semantic analysis, information extraction, information retrieval, machine translation and the like.