The present study aimed to explore the association between dietary patterns in abdominal obesity obtained by reduced-rank regression (RRR) with visceral fat index (VFI) as a dependent variable and dyslipidemia in rural adults in Henan, China. A total of 29538 people aged 18–79 were selected from the Henan Rural Cohort Study. RRR analysis was used to identify dietary patterns; logistic regression analysis and restricted cubic spline regression models were applied to analyze the association between dietary patterns in abdominal obesity and dyslipidemia. VFI was used as a mediator to estimate the mediation effect. The dietary pattern in abdominal obesity was characterized by high carbohydrate and red meat intake and low consumption of fresh fruits, vegetables, milk, etc. After full adjustment, the highest quartile of dietary pattern scores was significantly associated with an increased risk of dyslipidemia (OR: 1·33, 95 % CI 1·23–1·44, Ptrend < 0·001), there was a non-linear dose–response relationship between them (Poverall-association < 0·001, Pnon-lin-association = 0·022). The result was similar in dose-response between the dietary pattern scores and VFI. The indirect effect partially mediated by VFI was significant (OR: 1·07, 95 % CI 1·06–1·08). VIF explained approximately 53·3 % of odds of dyslipidemia related to the dietary pattern. Abdominal obesity dietary pattern scores positively affected VFI and dyslipidemia; there was a dose-response in both relationships. Dyslipidemia progression increased with higher abdominal obesity dietary pattern scores. In addition, VFI played a partial mediating role in relationship between abdominal obesity dietary pattern and dyslipidemia.