ZigBee is the primary standard solution for wireless sensor networks, implementing the Ad hoc On-Demand Distance Vector (AODV) protocol in the network layer and supported by the standard IEEE 802.15.4. This study is focused on mesh topologies and the critical problems encountered when AODV is executed in conjunction with the Carrier Sense Multiple Access with Collision Avoidance protocol. These problems are mainly related to the packet overhead required to carry out route creation. To perform preliminary experiments to be able to implement AODV in a real network, a new metric is proposed herein. This metric uses fuzzy logic to help in the decision-making process. The objective of the fuzzy routine is to determine, during the route-discovery process, the best node to forward request/reply packets, with the aim of reducing packet overhead and energy consumption. Moreover, minor changes are also added to the discovery procedure of AODV to improve the performance of the route-creation process. This intelligent version of AODV has provided promising experimental results, greatly reducing the number of packets required, with the consequent energy saving while selecting the best nodes to be part of the routes.