In Japan, large-scale epidemiological studies on starch and sugar intake are scarce, mainly due to a lack of a suitable assessment tool. We examined the relative validity of two widely-used dietary assessment questionnaires for Japanese adults, the comprehensive Diet History Questionnaire (DHQ) and the brief DHQ (BDHQ), for estimating the intake of starch and 10 types of sugars: total sugar, sucrose, maltose, lactose, trehalose, glucose, fructose, galactose, and added and free sugars. A total of 92 women and 92 men completed 4-day weighed dietary records (DRs) besides the DHQ and BDHQ in each of the four seasons. For each method, starch and sugar intake was calculated according to a recently developed food composition database on starch and sugars for Japanese food items. For most of the carbohydrate variables examined, the median energy-adjusted intake derived from the first DHQ and BDHQ (DHQ1 and BDHQ1, respectively) significantly differed from those derived from the 16-day DRs in both sexes. Spearman correlation coefficients between the 16-day DRs and DHQ1 were acceptable (≥0.31) for all variables (0.31-0.67), except for maltose and trehalose in women (≤0.29). For BDHQ1, the correlations were also acceptable for all variables (0.32-0.64), except for maltose (≤0.26) and galactose (≤0.06). Similar results were observed for the mean of four DHQs and BDHQs. This study indicated a reasonable ranking ability of DHQ and BDHQ for the intake of starch and most sugars examined, despite a poor ability to estimate the intake at the both group and individual levels.