<p>The explosive growth of Internet of Things (IoT) services has made it more challenging to handle the enormous volume of tasks produced by smart devices. Resource-constrained IoT devices often struggle to process these tasks locally, necessitating task offloading to external computational resources. This paper proposes an innovative dynamic task management approach using optimization algorithm (DTMO) for task offloading and scheduling. For computational offloading, edge server placement is optimized using the fuzzy C-Means (FCM) algorithm, thereby minimizing latency and improving communication between IoT devices and edge servers. Tasks from IoT devices are prioritized based on their urgency and computational requirements and are either processed locally or partially offloaded to edge servers using the consistency-based priority assignment (CPA) approach. A load-balancing strategy is also implemented to evenly distribute tasks between multiple edge servers, thus preventing bottlenecks and enhancing overall system performance. Task scheduling within virtual machines (VMs) on the edge servers are optimized using the self-improved rhinopithecus swarm optimization (SI-RSO) algorithm, which accounts for factors such as reliability, communication cost, resource utilization, and energy consumption, thereby ensuring optimal network Performance. The SI-RSO strategy reduces energy consumption by 32.61%, communication cost by 25.18%, execution time by 51.99%, processing time by 51.44%, and makespan by 37.40%, surpassing the conventional methodologies. In addition, SI-RSO enhances system reliability to 71.45% and achieves a resource utilization of 89.06%. This approach offers a scalable and efficient solution for managing IoT workloads, making it well-suited for diverse applications such as smart cities and industrial automation.</p>