Doctors have been troubled by false answers given by patients when describing their smoking, drinking, or medication behaviors as these false answers can disrupt effective treatment. Therefore, this study aimed to investigate false responses about smoking in medical examinations of university students. At their yearly medical examination, university students were required to answer questionnaires about their smoking habits. We examined 20,514 undergraduates’ questionnaires at a certain university from 2008 to 2013. The smoking situation was described using three phases: “never”, “past”, and “now”. We regarded students answers of “never” after previously answering “past” or “now” as false answers. Among the men, 6.7% of all answers given were false, whereas 18.5% of all answers given by the women were false. The low smoking rate of acquaintances led to students’ giving false answers, and students who gave a false answer once tended to repeat it. We suggest that measures implemented to improve unhealthy lifestyles should be considered and that those measures should be based on the finding that the actual proportion of persons leading an unhealthy lifestyle was greater than that indicated by the answers to questionnaires about lifestyle.