Abstract: Background: In most reactive routing protocols, the discovery process for any dictation starts by sending a RREQ (Route request) to neighboring nodes. Then, the intermediate nodes will rebroadcast the RREQ until the destination is found. Objective: In this study, new techniques have been suggested to reduce the broadcast of the RREQ packets in the network, when the reactive routing protocols are used. These techniques are Closest Candidate Neighbors for Rebroadcasting the RREQ (C-CNRR) and Furthest Candidate Neighbors for Rebroadcasting the RREQ (F-CNRR). Methodology: The key concept behind these two routing protocols is to divide the transmission range for each node that needs to find a path to a specific destination to four equal zones. Then, if no route is found in its routing table, one node per zone will be smartly selected to rebroadcast the received RREQ. Results: The simulation results show that in term of overhead, the C-CNRR and F-CNRR have better performance than the traditional ad hoc on-demand distance vector (AODV) routing protocol. Conclusion: Furthermore, reducing the overhead by more than 15% with our proposed techniques is reflected in an increase in the network throughput and a diminishing of data dropping.