The Internet of things (IoT) is a concept that allows all things such as sensors, actuators, humans, platforms, and software to interact with others relying on Internet infrastructure. The launch of the 5th generation communication networks allows providing network services with extremely high performance aiming to realize IoT concepts and forming a series of other smart IoT ecosystems. Nowadays, IoT is emerging as a dominant technology in almost domain. One of the most crucial technologies forming IoT networks is device-to-device (D2D) communication. D2D technology allows mobile nodes self-setup, and self-config to identify optimal paths for data transmission without relying on pre-installed infrastructure, which is the principle of forming mobile ad-hoc networks (MANETs). The fusing of MANETs and IoT technologies forming MANET-assisted IoT networks architecture to provide abilities and novel unpredict services. However, mobility and network structure often change are the main limitations of this architecture, hence, identifying the optimal route is a significant challenge. In this study, we propose a cross-layer routing approach for MANET-assisted IoT networks. Our routing mechanism is based on considering three factors, including number of hops, link quality, and queue length. The simulation results show that our proposed solution improves performance compared to the traditional solutions. We hope that our simple and efficient solution will be applied popular for future MANET-assisted IoT networks.