In this paper we are introducing a new method of tree formation, we propose a coordinate based method by which we can store and access tree structures. As we know in NLP, parsing is the most important module. The output of this module is generally parsed trees. Currently, TAG (Tree Adjoining Grammar) is widely used grammar due to its linguistic and formal nature. It is simply tree generating system. The unit structure used in TAG is structured trees. So we used our new method to store trees where we worked on English to Hindi language. We worked on different sentences from English to Hindi, our method is the easiest way to manipulate tree. We have implemented within small corpus and for finite number of structures and further can be extended in future.