This paper proposed an algorithm based on the improved Hidden Markov Model (HMM) to achieve the Chinese Named Entity Recognition (NER) in Nature Language Processing (NLP) .It put forward three kinds of familiar recognition methods on the basic of summarizing the trait and difficulty of Chinese NER. The application of the Hidden Markov Model (HMM) based on statistical method in NER was researched. It pointed out and analyzed the limitation of the conventional HMM. At the same time, some improvements have been made in addition.