Abstract: To deeply investigate the decision making for microscopic lane-changing behaviors on freeway, and to improve the authenticity and effectiveness of corresponding lane-changing models, discretionary lane-changing behaviors were divided into leftward and rightward lane-changing behaviors in this research; with the factors that impact drivers' characteristics and vehicle type parameters introduced into the research, a discretionary lane-changing model was established based on fuzzy logic. Meanwhile, vehicle speed control models under the free state and vehicle following models were built to enhance dynamic effects of lane-changing behaviors. The simulation results suggest that the discretionary lane-changing model is highly authentic and valid. According to the research, both the leftward and rightward lane-changing models can accurately reflect the decision making for microscopic lane-changing behaviors on freeway.