Application of variable speed-limit (VSL) control has become promising in managing the incidents in the contexts of intelligent highway systems wherein advanced VSL signs are densely placed along the roadways. This paper aims to develop a genetic-fuzzy-logiccontroller (GFLC)-based VSL control model with an objective function to enhance both safety and throughput efficiency at the incident sites. In order to evaluate the performance of learned logic rules and tuned membership function, a revised cell transmission model (CTM) is proposed, which is validated with field traffic data. The results indicate that the proposed VSL control model can effectively curtail the crash likelihood without deteriorating or only slightly decreasing the throughput efficiency under different tested incident cases. It suggests that our proposed GFLC-based VSL control model is promising for practical applications.