Discourse analysis is an important natural language processing task. There are many discourse parsers in many languages, such as English and Chinese, constructing discourse trees from text documents for further semantic analysis. However, there is no official release of Vietnamese discourse treebank for research in Vietnamese discourse parser. Therefore, this paper presents our preliminary result in building Vietnamese discourse treebank. some problems when building discourse treebank and proposes a discourse annotation framework for it. In order to show the feasibility of developing discourse parsers for Vietnamese documents, two experiments in discourse relation classification and in discourse nucleus classification are conducted using the discourse annotated documents.