in cross-language information retrieval (CLIR), the quality of retrieved result from the information retrieval process depends on a user's query. by supposing query as set of keywords, CLIR researchers in various languages focus on the automatic proces of translating and disambiguting queries. many of them relate to transliton from chinese to english and select the possible results. the solution can apply not only for vietnamese, but also for other languages. in details, with the suport of a special bilingual dictionary, the procesing model works in a pipeline to translate the key-phrases of source query into corresponding key-phrases in target laguages. furthermore, basing on this approach, the CLIR system may retrieve the possible context and improve the retrievend results. regarding the processing model, the paper also proposes algorithms to process the user's entered key-phrases by using the interative proposing. the paper may not cover all of CLIR cases; however it shows an algorithm that provides the good result and high accuracy with sample test data in vietnamese and english.