In this paper, we wrote some rules of morphology and syntax oriented JavaCC in order to make use of JavaCC parser and implemented the application in Japanese-Mongolian machine translation. Firstly, by analyzing the morphological features of Japanese language and then doing syntactic parsing based on chunks, lastly, completing the transformation and generation of lexicon and structure. A test of parsing 1519 sentences from 310 Japanese newspaper reports resulted in the accuracy of 100% and another one test of translating 2183 chunks in 50 Japanese reports showed the accuracy of 95.37%.