Aims The aims of this study were to evaluate the associations of metabolic abnormalities with incident diabetic kidney disease (DKD) and to explore whether dyslipidaemia, particularly high fasting triglyceride (TG), was associated with the development of DKD. Methods In total, 11 142 patients with new-onset type 2 diabetes with baseline estimated glomerular filtration rates (eGFR) >= 60 mL/min/1.73 m(2) were followed up during 2011-2016. Incident DKD was defined as eGFR <60 mL/min/1.73 m(2) at follow-up. Multiple logistic regression analysis was conducted to explore the relationship of metabolic abnormalities at baseline and at follow-up with risks of DKD. High TG was defined by TG >= 1.70 mmol/L. Low high-density lipoprotein cholesterol (HDL-c) was defined by HDL-c Participants who developed DKD had higher levels of waist circumference and systolic blood pressure, and lower levels of HDL-c at both baseline and follow-up visits. The DKD group also had higher levels of post-load plasma glucose and TG at follow-up. Multivariate logistic regression analysis revealed that both high TG at baseline [odds ratio (OR) = 1.37, p = .012) and high TG at follow-up (OR = 1.71, p < .001) were significantly associated with increased risks of DKD. Patients with high TG levels at both baseline and follow-up had higher risk of DKD compared with constantly normal TG (OR = 1.65, p < .001) after adjustment for covariates. Conclusions In a large population of patients with new-onset type 2 diabetes, a high TG level was an independent risk factor for the development of DKD. Tight TG control might delay the occurrence of DKD.