At present, the research on machine translation mainly focuses on English-Chinese translation, while the research on Japanese college students using Japanese Chinese machine translation software is relatively few. In order to solve the above problems in Chinese Japanese bilingual translation, this paper proposes a phrase translation method based on sequence intersection. This method regards sentences as word sequences and aligns the sequence intersection of all source sentences corresponding to the target sentence in the corpus with Chinese and Japanese sentences containing the phrases to be translated. By fully mining the information of sentence alignment bilingual corpus without word alignment resources, we can obtain high-quality phrase translation, syntactic analysis, and dictionary. Then, we focus on the automatic construction of sentence level aligned bilingual corpus and explore the automatic sentence alignment technology of Chinese and Japanese bilinguals. A ten-year alignment model based on combination cues and core extended square matching is proposed. The preprocessing of the computer corpus and the basic construction of the corpus are completed. This paper also puts forward corresponding countermeasures and approaches to the problems encountered in the construction of computer translation.