Abstract Background The Health Literacy Questionnaire (HLQ) is an increasingly used health literacy instrument that has been translated into many languages. The HLQ has 44 items and comprises 9 scales assessing the multidimensional construct of health literacy. This study reports the HLQ reliability and construct validity tested in people with chronic diseases living in Vietnam. Methods Adults (n = 600) hospitalized with chronic disease in Vietnam completed the HLQ. Floor and ceiling effects, item, and scale difficulty levels were assessed. Generalized linear models with backward modeling techniques were performed to test key variables associated with each HL domain. Confirmatory factor analyses (CFA) testing nine one-factor models were fitted to test the structure of each scale, and a nine-factor model tested the hypothesized structure of the HLQ, followed by the calculation of scale reliability using Cronbach’s alpha. Results No item had floor effects, and only eight items showed ceiling effects. Two scales that had the most difficult tasks to complete (highest difficulty level) were 8 “Ability to find good information” and 9 “Understanding enough to know what to do”. Variables associated with health literacy were education, income, age, residential area, main support persons and comorbidity index (associated with 7, 7, 4, 3, 2 and 2 out of 9 scales, respectively). Each HLQ scale demonstrated a robust unidimensional construct with all CFI ≥ 0.95, RMSEA varied from 0 to 0.07. The nine-factor CFA model demonstrated satisfactory fit indices: X2 = 5537.4, 866 df, p