The purpose of this study is to investigate the risk classification factor in motorcycle compulsory liability insurance, and analyzes it by Logistic regression. Expect for the engine displacement which the insurance companies has adopted as the risk factor now, this research further examines the factors such as: gender and age, car age, car brand, and regional factor .This paper indicate that both the human factors and the vehicle factors, and further adopt bonus-malus system are used to make the risk classification more perfect. This study provides the suggestion of a better risk classification for motorcycle in compulsory liability insurance.