Load balancing service is essential for distributing workload across server farms or data centers and mainly provided by dedicated hardware. In recent years, the concept of Software-Defined Networking (SDN) has been applied successfully in the real network environment, especially by OpenFlow designs. This paper presents an OpenFlow-based load balancing system with the genetic algorithm. This system can distribute large data from clients to different servers more efficiently according to load balancing policies. Furthermore, with the pre-configured flow table entries, each flow can be directed in advance. Once the traffic burst or server loading increased suddenly, the proposed genetic algorithm can help balance workload of server farms. The experiments demonstrate the better performance of the proposed method compared to other approaches.