According to the studies about the second language learning, learners tend to use his first language to emulate the pronunciation of the second language. The language pronunciation problem caused by using the first language is regarded as ”pronunciation obstacle.” In this paper, the pronunciation obstacle of British and American to the study of Mandarin is explored. Based on the corpus, we will propose a pronunciation obstacle identification system on learning Mandarin. This system will be useful for learners to have more native pronunciation based on the information provided by this system.