Software-based load balancers implemented via Software-Defined Networking (SDN) offer cost and configuration advantages over hardware solutions in Cloud environments. However, existing load balancing methods in the OpenStack platform lack dynamic approaches, relying primarily on static mechanisms. This work addresses the limitation by proposing a dynamic load balancing algorithm, DWRR (Dynamic Weighted Round Robin), based on CPU usage metrics. An agent implementing DWRR was successfully deployed in a cloud environment using OpenStack on the VirtualBox hypervisor. The performance of DWRR was compared against the static WRR (Weighted Round Robin) algorithm in allocating workloads across web servers based on CPU utilization. Results demonstrate DWRR’s superiority in efficiently distributing workloads by considering real-time resource usage metrics, leading to improved load balancing and resource utilization in the cloud environment. The proposed dynamic approach outperforms static load balancing methods, addressing a critical issue in Cloud computing.