In recent years, language model has been utilized language and speech processing, such as speech recognition, speech synthesis, information retrieval, machine translation, optical character recognition and so on. Since, natural language is highly diversified, new words, e.g., cloud computing, and new phrases are created constantly. Moreover, in different areas, different time or different circumstances, same words or same phrases may have different semantic meaning. Because of these varieties, language model adaptation is becoming more and more important. In this paper, we will introduce some techniques of language model adaptation and give a detail analysis and comparison.