機車具有便利、快捷與機動等多項功能，在青少年族群中代表著不同的意義，然 而由於年齡的限制，所以無照騎車就很容易發生，尤其在大眾運輸系統較差、警察取締較鬆的地區。同時，無照騎車又比其他交通違規行為更具危險性，隨時都有可能發生嚴重、致命的交通事故傷害。因此，探討青少年無照騎車行為之影響因素，就成為防止他們機車事故傷害非常重要的工作。然而，由於目前國內、外探討青少年無照騎車行為影響因素之研究仍待累積，所以本研究即以修改後健康信念模式為理論架構，應用結構方程模式來探討青少年無照騎車行為之影響因素。結果發現此模式對於青少年無照騎車具有良好的預測力，並且自我效能、安全價值、有效性與障礙性認知是行為動機與行為預期之重要影響變項，而行為動機、行為預期、自我效能與安全價值則是行為之重要影響變項。同時，此模式各主要變項之重要信念均能有效的區分青少年有沒有無照騎車行為，這可作為未來發展青少年安全騎車行為之教育計畫的重要內容。再者，此模式對行為的預測不會因性別、年級、年齡不同而有所差別，然而過去行為則可影響此模式對行為的預測，所以此模式更適合於預測未曾有過無照騎車者他們未來的行為。

　　　　　The motorcycle is convenient, speedy, and movable, and it does represent the particular meaningfulness in the adolescents; however, it often occurs to the adolescent motorcyclists without license under the limit of driver age, especially in the poor public transportation system and the loose traffic police check-up. It is more dangerous to drive motorcycles without license than to commit other traffic violations because the former may cause serious and fetal traffic injuries. Therefore, it is very important to explore the influential variables of the adolescent motorcyclists without license for preventing them from the motorcycle injuries. There are very few studies for the adolescents' driving motorcycles without license whether domestic or foreign. The modified health belief model (MHBM) is taken as the framework of this study, and the structural equation models are used to understand the influential factors of the adolescent motorcyclists without license. The following results are found. The model offers the good prediction for the adolescents' driving motorcycles without license. Self-efficacy, safety value, perceived benefits, and barriers are four important variables of the behavioral motivation and expectation. Behavioral motivation, behavioral expectation, self-efficacy, and safety value are the important variables of the behavior in driving motorcycles without license. Again, the significant beliefs of the principal components in MHBM can efficiently discriminate the adolescents' driving motorcycles without license. It can offer us to develop the education program for the adolescents' driving motorcycles safely. Furthermore, the prediction of the model on the behavior is not different for gender, grade, and age, instead of the past behavior. The prediction of the model on the behavior is better for not driving than for driving motorcycles without license.