WSN consists of nodes that are very difficult to replace by nature due to their physical placement in remote and inaccessible locations. These nodes are equipped with limited energy resources. When the nodes perform their tasks, they consume power. Thus, power consumption becomes a priority in the implementation of WSN. In order to manage power consumption for better network performance and node lifetime, many wireless sensors network routing protocols are used. The Geographical Adaptive Fidelity (GAF) protocol is one of the best location-based routing protocols. This protocol breaks down the network into a virtual grid and considers grids as clusters, with the node with the highest energy selected as the cluster head for data transmission. Choosing only the one with the highest power cannot reduce energy waste or prolong the network's lifespan, as nodes with lower power die without being used. In this thesis, we present a modified version of GAF, i.e., MGAF, which is proposed to improve cluster head selection techniques to minimize energy utilization and extend network lifetime. The proposed Modified GAF algorithm is implemented in MATLAB. Two experimental simulation cases have been conducted. Ten simulations were carried out in each case. In the first and second experimental simulation cases, modified GAF saved 59.5% and 107.75% of energy, respectively. And also, the result of modified GAF shows higher throughput and fewer routing overheads than basic GAF. Therefore, modified GAF significantly minimizes energy utilization and increases network lifetime and performance compared to basic GAF. Keywords- Energy Efficiency, GAF, Modified GAF, WSN, GH