Auxiliary English Writing Method Based Chinese Pinyin

English writing assistant method like input method is one of main researches to improve English writing efficient. How to make Chinese natives write English more conveniently,quickly and naturally is the key problem of this research issue to solve.Current English writing assistant systems are mainly based on dictionary matching and sentence database retrieval, it is a traditional way which is limited by the user's English language proficiency while the assistant efficient is low. Current Chinese writing technology especially input method not only enable users to input pinyin with high writing interaction efficiency, but also recommends words with a high accuracy. On the other hand, machine translation can be seen as conversion process between two different codes, phrase translation model trained by phrasal-based statistical machine translation system(PBSMT) gives the statistical phrase translation probability of the two different codes. This paper analyzes main problems like large information and difficult to determine a word encountered by non-natives in writing process, and proposes an English writing assistant method based Chinese pinyin drawing lessons from the convenience input method. This writing assistant method implements a cross language input approach which gives recommendation of English words and phrase through inputting Chinese pinyin when a user do not know how to express a difficult English word without frequent switching between tools like dictionary and the text edition tool.The main work include: Firstly, detect simplified pinyin spelling codes like jianpin or incomplete quanpin user input in the process of interaction and get complete pinyin string through pinyin segmentation and incompletion. Secondly, establish personalized pinyin-English word library, do fuzzy matching of pinyin codes with the library to get candidate words, then sort and disambiguation them by word frequency, language model and.Finally train pinyin-English phrase static machine translation system and give relative phrase or words conjunction based on the phrase table. We use recommendation accuracy of difficultwords in the writing interaction process as evaluation means, and the result verified the availability and reliability of this system itself.