Recent studies indicated that vitamin A (VA) might be involved in the pathology of type 2 diabetes mellitus (T2DM). This cross-sectional study was conducted to explore the association between circulating VA level and T2DM. A total of 1 818 subjects aged 50 years old and above were recruited from the community. Binomial logistic regression and restricted cubic spline (RCS) were applied to analyze the association of plasma VA level with the risk of T2DM. Serum VA and lipid-adjusted VA levels of T2DM patients were significantly higher than that of non-T2DM subjects (P < 0.05). The ratios of plasma VA/total cholesterol (TC), VA/high-density lipoprotein cholesterol (HDL-c) and VA/low-density lipoprotein cholesterol (LDL-c) were positively associated with the risk of T2DM in the aging population (P < 0.05). Compared with the Q1 level, subjects with Q2 to Q3 levels of plasma VA/triglyceride (TG) have decreased risk of T2DM (odds ratio (OR)Q2 = 0.68, PQ2 = 0.021; ORQ3 = 0.59, PQ3 < 0.01). Our results indicated that the imbalance of circulating lipids and VA might affect the relationship between VA and T2DM. The middle and aging subjects with higher ratios of plasma VA/TC, VA/HDL-c, and VA/LDL-c displayed increased risk for T2DM, but the moderate ratio of VA/TG might protect against risk of T2DM.