Novel dietary assessment methods are needed to study chronic disease risk in agrarian cultures where food availability is highly seasonal. In 16,320 rural Nepalese women, we tested a novel food frequency questionnaire, administered once, to assess past 7-day intake and usual frequency of intake throughout the year for year-round foods and when in season for seasonal foods. Spearman rank correlations between usual and past 7-day intakes were 0.12–0.85 and weighted kappa statistics, representing chance-corrected agreement, were 0.10–0.80, with better agreement for frequently consumed foods. The questionnaire performed well, but may require refinement for settings of extremely low dietary diversity.