Providing efficient encoding approaches for mining XML query patterns is crucial, as many applications use XML to share data in their disciplines over the Internet. These encoded XML query patterns can be used to design an index mechanism or cached and thus enhance XML query performance. Several XML query pattern mining algorithms have been proposed to record user queries in the system and thus discover the frequent XML query patterns. By using the frequent XML query patterns, the query performance of XML data is improved. However, to find out the frequent XML query patterns from big XML data over the Internet, these existing algorithms are time-consuming and thus are not suitable. MapReduce is a software framework for writing parallel programs and can be used to enhance the performance to encode XML query patterns. In this paper, MapReduce-based algorithms to parallel encode XML queries, report experimental results of the prototype implementation on Hadoop system, and a popular implementation of MapReduce framework are proposed.