Although the Japanese EuroQol instrument was developed in 1998 by a Japanese translation team, the reliability of health state valuations of the fuller questionnaire has not yet been investigated in Japan. We carried out a pilot study to investigate the reliability of health state valuations of the Japanese EuroQol instrument.The survey was conducted twice with the fuller questionnaire of the Japanese EuroQol instrument in a class of preventive medicine. We analyzed test-retest reliability based on the health state valuations and calculated Cronbach's coefficient alpha of both tests.The mean valuations of 14 hypothetical health states and 'death' between test and retest were not significantly different by paired t-test. The reliability by calculated correlation between test and retest was 0.996 (p<0.0001). In addition, Cronbach's coefficient alpha of the first test was 0.827, and that of the second test was 0.865. Although good reliability was shown in the present study, our conclusion was limited to applications of population-based surveys because of the small number and limited subjects of the present study. Further investigations are required in the form of a population-based survey.