Searching refers to the process of finding a data value within some given set of data values in the form of list or array or any other form. Searching has obtained so much concern on a global scale the reason being the increasing amount of data day by day so there is a need to reduce the searching time and thus we require efficient searching methodology. Taking the binary search as a base for the searching technique goal is to find a required element within an ordered list by making comparisons. Size of list reduces after each iteration. This paper proposes a new algorithm for searching. The main point of difference is that it uses variable partitioning within the list for the first three iterations and again start repeating the same procedure. Binary Search can be improved by increasing the number of sublists but doing so will increase the number of comparisons thus not a good idea. The proposed algorithm overcomes this drawback of breaking into more sublists by variable partitioning.