To evaluate various diet quality indices and to estimate their associations with major non-communicable diseases (NCD) (i.e. diabetes mellitus (DM) and myocardial infarction (MI)) and risk for overweight (OW).

Four dietary diversity indices (namely, count index (Count), dietary diversity score index, berry index (BI) and entropy index (EI)) and three Chinese dietary guideline-based indices (namely, China healthy diet index, Chinese food pagoda score and diet quality divergence index) were employed to evaluate Chinese diet quality. DM, MI and OW were used as diet-related health indicators. Logit regressions were employed to unveil the associations between diet quality indices and NCD and risk for OW. The relationships between diet quality indices and daily energy intakes were checked with ordinary least squares linear regressions.

Four recent waves (2004, 2006, 2009, 2011) of longitudinal individual data from China Health and Nutrition Survey.

Chinese adults (aged 18–64 years) from twelve provinces were included in the analysis (n 30 350).

Count, BI, and EI were positively associated with higher OW risk and daily energy intakes. As dietary guideline-based indices got better, people were exposed to lower DM and OW risks and got lower daily energy intakes. Finally, dietary guideline-based indices properly revealed the expected relationships that high-quality diets would reduce NCD and risk for OW, while high diversity indices were usually correlated with over-nutrition and high risks.

Increasing diversity of the diet does not necessarily improve the nutrition and health. Dietary guideline-based indices are more robust than dietary diversity indices; thus, they should be highly recommended when evaluating diet quality.