There is increasing evidence for the role of nutrition in the prevention of depression. This study aims to describe changes in diet quality over 12 years among participants in the Australian Longitudinal Study on Women’s Health in relation to changes in depressive symptoms. Women born between 1946 and 1951 were followed-up for 12 years (2001–2013). Dietary intake was assessed using the Dietary Questionnaire for Epidemiological Studies (version 2) in 2001, 2007 and every 2–3 years after that until 2013. Diet quality was summarised using the Australian Recommended Food Score (ARFS). Depressive symptoms were measured using the ten-item Centre for Epidemiologic Depression Scale at every 2–3-year intervals during 2001–2013. Linear mixed models were used to examine trends in diet quality and its sub-components. The same model including time-varying covariates was used to examine associations between diet quality and depressive symptoms adjusting for confounders. Sensitivity analyses were carried out using the Mediterranean dietary pattern (MDP) index to assess diet quality. Minimal changes in overall diet quality and its sub-components over 12 years were observed. There was a significant association between baseline diet quality and depression (β=−0·24, P=0·001), but this was lost when time-varying covariates were added (β=−0·04, P=0·10). Sensitivity analyses showed similar performance for both ARFS and MDP in predicting depressive symptoms. In conclusion, initial associations seen when using baseline measures of diet quality and depressive symptoms disappear when using methods that handle time-varying covariates, suggesting that previous studies indicating a relationship between diet and depression may have been affected by residual confounding.