One of the major challenges in Mobile Ad Hoc Networks (MANETs) is achieving Quality-of-Service (QoS) for data communication using multi-disjoint routing protocols. As a result, a Reliable and Stable Topological change Adaptive Ad Hoc On-demand Multipath Distance Vector (RSTA-AOMDV) routing protocol has been designed to select forwarding nodes based on the node's local information. However, this protocol did not consider the node's global stability, and predicting network topology in high-speed MANETs was difficult due to the random movement of nodes. Therefore, this article proposes a novel routing protocol to choose stable nodes and optimal routes for effective data transmission in high-speed MANETs. First, an Energy-conserved Stability and Density-aware QoS-enabled Topological change Adaptable Multipath Routing (ESDQTMR) protocol is developed to ensure data transmission stability from the source to the target nodes by considering both local and global stability factors. Stable Betweenness Centrality (SBC) and transferring packets according to the locality dependency energy level are adopted to find a stable node. After that, an Optimized ESDQTMR (OESDQTMR) protocol is proposed, which uses the Golden Eagle Optimization (GEO) algorithm to choose the best path according to the new objective function. Additionally, fractional calculus is considered to enhance the exploration of the optimal routing path. Finally, extensive simulation findings show that the OESDQTMR protocol achieves greater network performance compared to other routing protocols under varying numbers of nodes and node mobility.