The rapid growth of Internet of Things (IoT) devices and the deployment of unmanned aerial vehicles (UAVs) as aerial base stations (BS) introduce significant challenges in uplink random access for massive machine-type communication (mMTC) networks. This paper proposes a Q-learning-based random access scheme integrated with Sparse Code Multiple Access (SCMA) tailored for UAV-assisted IoT scenarios. By enabling devices to autonomously optimize time slot group selection and codebook allocation through Q-learning, the scheme enhances access success probability and throughput while satisfying practical system constraints. A customized reward function ensures stable learning convergence and efficient resource utilization. Simulation results demonstrate the effectiveness and robustness of the proposed approach in improving access performance under dense device deployments, validating its suitability for scalable and reliable uplink access control in UAV-enabled mMTC networks.