<p>In fog computing, devices located at the edge of the network are equipped with computational and storage resources to provide services closer to users. The main goal of fog computing is to reduce the latency of requests, particularly those that are sensitive to delay, and to decrease the computational load on centralized data centers. Task scheduling is a major challenge in achieving these objectives, as it directly affects the performance of the fog-cloud infrastructure and must therefore be carefully addressed. This paper presents a Reinforcement Learning-based Task Scheduling (RLTS) approach for fog-cloud environments. First, the Best-Worst Method (BWM) is used to determine the priority of tasks. Then, Reinforcement Learning (RL) is applied to allocate tasks to appropriate nodes. In the first stage, high-priority tasks, which are sensitive to delay, are assigned to suitable fog nodes. In the second stage, low-priority tasks, which can tolerate delay, are allocated to cloud nodes. The proposed scheduling algorithm aims to increase the percentage of tasks that meet their deadlines while reducing system load and waiting time. Compared to existing scheduling algorithms, the proposed method improves task completion rates and reduces load, response time, energy consumption, and waiting time.</p>