We evaluated the reproducibility and validity of an FFQ used in a Chinese community-based nutrition and health survey. A total of ninety-nine males and 104 females aged 31–80 years completed four three consecutive 24-h dietary recalls (24-HDR, served as a reference method, one three consecutive 24-HDR for each season) and two FFQ (FFQ1 and FFQ2) over a 1-year interval. The reproducibility of the FFQ was estimated with correlation coefficients, misclassification and weighted κ statistic. The validity was evaluated by comparing the data obtained from FFQ2 with the mean 24-HDR (m24-HDR). Compared with the m24-HDR, the FFQ tended to underestimate intake of most nutrients and food groups. For all nutrients and food groups, the Spearman’s and intra-class correlation coefficients between FFQ1 and FFQ2 ranged from 0·66 to 0·88 and from 0·65 to 0·87, respectively. Most correlation coefficients decreased after adjusting for energy. More than 90 % of the subjects were classified into the same or adjacent categories by both FFQ. For all nutrients and food groups, the crude, energy-adjusted and de-attenuated Spearman’s correlation coefficients between FFQ2 and the m24-HDR ranged from 0·21 to 0·69, 0·19 to 0·58 and 0·25 to 0·71, respectively. More than 70 % of the subjects were classified into the same and adjacent categories by both instruments. Both weighted κ statistic and the Bland–Altman plots showed reasonably acceptable agreement between the FFQ2 and the m24-HDR. The FFQ developed for adults in the Nanjing area can be used to reliably and validly measure usual intake of major nutrients and food groups.