The optimization of spatial indexing is an important issue considering the fact that spatial database, in such diverse areas like geographical, CAM and image applns are growing rapidly in size and often contain in the order of millions of items.To handle these multi-dimensional data, R-tree is widely used as data structure. The node splitting algorithm used in R-tree process affects the query performance and results in an inefficient R-tree structure as it generates uneven nodes. To overcome these drawbacks, we have proposed an algorithm to balance the uneven node splitting to meet the demand of the R-tree process. The projected algorithm inserts the node into the sibling instead of splitting or re-insertion of the overflow node which paves way to reduce the overhead of splitting process, adjusting tree construction operation and number of disc accessing.