Cancer education programs are regularly conducted in schools in Japan. Previous reports on their effectiveness were based on surveys conducted several months after the education. We aimed to evaluate whether cancer education for children influenced their behavioral changes in adulthood. We targeted schools where the Japan Cancer Society has conducted the programs since 2011 and the program participants are older than 21 years currently. Invitations were sent to the participants and controls who graduated one year before or after the participants, and answers were obtained using an online questionnaire. Data were collected on the willingness to undergo cancer screening, sociodemographic characteristics, healthy behaviors, and health literacy. We compared the data from the program participants with those from controls. We requested cooperation from 13 schools; however, only one agreed. The common reason for refusal was privacy concerns. In the all-boys participating school in Tokyo, there was no significant difference in background information between the cancer education (38 participants) and control (55 participants) groups. Multiple linear regression showed that healthcare or welfare education background (β = 0.25, p = 0.01) and health literacy (β = 0.24, p = 0.02) were significant predictors of cancer screening intention, while the presence of cancer education (p = 0.25) was not. Despite severe selection bias, this is the first study to examine the long-term impacts of cancer education. We found no significant impacts on the measured outcome. However, the educational content at that time differed from that of today, and the program's efficacy should not be negated.