提出在基于实例的英汉机译中，一种基于相似性评估的近似实例求解策略·首先，对知识源进行预处理，包括树库、片段库及片段组合库的构建；其次，对输入句子进行预处理，包括句法分析树的建立，合法片段的抽取及片段组合形式的转换；最后，测量片段至句子的相似性，依据相似性阀值获取输入句子的所有近似实例·实验结果表明，各项测试指标都比较令人满意 In this paper, we propose an approximate example solving strategy based on similarity assessment in case-based English-Chinese machine translation. Firstly, the knowledge source is preprocessed, including the construction of tree library, fragment library and fragment combination library. Secondly, Pretreatment, including the establishment of syntax analysis tree, the extraction of legal fragments and the combination of fragments. Finally, the similarity between fragment and sentence is measured and all approximate examples of input sentence are obtained according to the similarity threshold. The experimental results show that, Test indicators are more satisfactory