语言模型是一种自然语言处理的工具，与信息检索相结合形成的LM-IR，是信息检索模型研究上的重大突破。而平滑化作为解决模型零概率问题的重要方法，在提高信息检索模型的性能方面具有重大的意义。本文介绍了基于文档长度信息检索的基本模型，重点介绍了语言模型信息检索模型的平滑化技术，最后从整体上介绍了平滑化方法的一些研究进展情况。

The language model is a kind of natural language processing tools. LM – IR ,which formed by the combination of language model and information retrieval, is a major breakthrough in the research on IR model. Smoothing is an important method to solve the problem of zero probability ,and it is significant to improve the performance of information retrieval model. This article introduces the basic model based on the length of the document information retrieval and focus on language model smoothing technology of information retrieval model. Finally, the article introduces some research progress of smoothing method .