Introduction: The aim of this study was to develop and validate a semiquantitative food frequency questionnaire (FFQ) for assessing habitual added sugar consumption of adults in the Klang Valley. Methods: In the development phase, a 24-hour dietary recall (24-hr DR) was used to determine food items to be included into the FFQ among adults from three major ethnicities (n=51). In the validation phase, the FFQ was further validated against a reference method which was a multiple-pass 24-hr DR among 125 adults in Klang Valley. The response rate for the latter phase was 96.1%. Results: The semi-quantitative FFQ consisting of 64 food items was categorised into 10 food groups. The mean added sugar intake determined by the reference method was 44.2 ± 20.2 g/day while that from the FFQ was 49.4 ± 21.4 g/day. The difference in mean intake between the two methods was 5.2 g (95% CI = 2.6-7.9; SD = 14.9, p<0.05) or 11.8%. Pearson correlation was r=0.74 (p<0.001) for the two methods while Spearman rank correlations for the various food groups ranged between 0.11 (cake and related foods) to 0.61 (self-prepared drinks), with most groups correlating significantly (p<0.05). Cross-classification of subjects into quintiles of intake showed 47.2% of the subjects correctly classifying into the same quintile, 34.4% into adjacent quintiles while none were grossly misclassified. The Bland-Altman plot was concentrated in the y-axis range (-24.14 g to 34.8 g) with a mean of 5.22 g. Conclusion: This semi-quantitative FFQ provides a validated tool for estimating habitual intake of added sugar in the adult population of the Klang Valley.