The study focuses on improving the Quality of Service (QoS) in Vehicle-to-Vehicle (V2V) communication within Vehicular Ad Hoc Networks (VANETs) by enhancing the Learning Automata-based Ad Hoc On-Demand Distance Vector (LA-AODV) routing protocol. Unlike the standard AODV, which is a reactive routing protocol, and previous configurations of LA-AODV, this research introduces a fine-tuning strategy for the learning automata parameters. This strategy allows the parameters to dynamically adapt to changing network conditions to reduce routing overhead and enhance transmission stability. Three modified versions of LA-AODV referred to as setups A, B, and C, are evaluated against the standard AODV and earlier LA-AODV configurations. The performance of each setup is measured using key QoS metrics: flood ID, packet loss ratio (PLR), packet delivery ratio (PDR), average throughput, end-to-end delay, and jitter. These metrics are crucial in evaluating the efficiency, reliability, and performance of V2V communication systems within VANETs. The results demonstrate that the LA-AODV variants significantly reduce flood ID counts, which represent the number of times a packet is broadcasted, compared to AODV, with setups A and B achieving reductions of 10.24% and 28.74%, respectively, at 200 transmissions, indicating enhanced scalability. Additionally, LA-AODV setup A provides 5.4% higher throughput in high-density scenarios. The modified versions also significantly decrease delay and jitter, achieving reductions of over 99.99% and 99.93%, respectively, at 50 transmissions. These findings underscore the adaptive capabilities of the proposed LA-AODV modifications, providing reassurance about the robustness of the system. They also highlight the importance of parameter optimization in maintaining reliable V2V communication. Future work will benchmark LA-AODV against other state-of-the-art protocols to validate its effectiveness further.