We examined the correlation between the prefectural variation of standardized mortality ratio (SMR) for lung and colon cancer and the consumption of foods using dietary data from National Nutrition Survey in Japan from 1976 to 1995. We found partly the same results on promoting factors and preventing factors of these cancers as the previously reported large scale cohort studies. But, we also found the inconsistent results. This might be partly because of the limitation of data collection method in the National Nutrition Survey. It was thought that the method of National Nutrition Survey should be modified to monitor nutrition and health conditions to correspond the epidemiological study and to prevent chronic diseases.