This thesis explores the types of mistakes trainees make during their class of interpretation of speakers whose mother tongues are Chinese dialects and the ways to solve the problems. Standard Mandarin which these speakers use is frequently interfered with by their dialect, particularly by the pronunciation rather than the vocabulary or sentence structure. Therefore, the mistakes these students make resulted from the speakers' pronunciation. They either omit or misinterpret when they find big differences in pronunciation between Mandarin and the speakers' dialects.In order to improve the ability to distinguish the pronunciation, the trainees need to correct the mistakes that they tend to make and prepare a data according to each dialect in addition to practicing interpreting the texts that these speakers use. To collect the data, the following three methods are effective: 1) the contrast and analysis of the dialect and Mandarin 2) the analysis of Mandarin of the speaker speaking a Chinese dialect 3) the analysis of the mistakes the trainees make.
