Abstract: Achieving the effectiveness in relation to the relevance of query result is the most crucial part of XML keyword search. Developing an XML Keyword search approach which addresses the user search intention, keyword ambiguity problems and query result grading problem is still challenging. In this study, a new keyword search approach, named XDMA (XML keyword search dual indexing and mutual summation algorithm), for XML databases based on dual indexing, is proposed to resolve these problems. This approach includes design of mutual score, entropy-based similarity measure so as to find the relevant results for a given XML keyword query and, grading score to grade the query results. The algorithm for the new approach is presented and finally, space and time complexity of the algorithm have been analyzed to show the effectiveness of the algorithm for XML keyword search. The proposed algorithm XDMA can retrieve the relevant results for XML keyword query more effectively.