Non-alcoholic fatty liver disease (NAFLD) is the most common liver disease. Phthalates have been suggested to influence the development of NAFLD due to their endocrine-disrupting properties, but studies based on nationally representative populations are insufficient, and existing studies seem to have reached conflicting conclusions. Due to changes in legislation, the use of traditional phthalates has gradually decreased, and the phthalates substitutes is getting more attention. This study aims to delve deeper into how the choice of diagnostic approach influences observed correlations and concern about more alternatives of phthalates, thereby offering more precise references for the prevention and treatment of NAFLD. A cohort of 641 participants, sourced from the National Health and Nutrition Examination Survey (NHANES) 2017-2018 database, was evaluated for NAFLD using three diagnostic methods: the Hepatic Steatosis Index (HSI), the US Fatty Liver Indicator (US.FLI), and Vibration Controlled Transient Elastography (VCTE). The urinary metabolite concentrations of Di-2-ethylhexyl phthalate (DEHP), Di-isodecyl phthalate (DIDP), Di-isononyl phthalate (DINP), Di-n-butyl phthalate (DnBP), Di-isobutyl phthalate (DIBP), Di-ethyl phthalate (DEP) and Di-n-octyl phthalate (DnOP) were detected. The association between NAFLD and urinary phthalate metabolites was evaluated through univariate and multivariate logistic regression analyses, considering different concentration gradients of urinary phthalates. Univariate logistic regression analysis found significant correlations between NAFLD and specific urinary phthalate metabolites, such as Mono-(2-ethyl-5-oxohexyl) phthalate (MEOHP), Mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), Mono-2-ethyl-5-carboxypentyl phthalate (MECPP), and Mono-(carboxyisoctyl) phthalate (MCiOP), across different diagnostic criteria. In a multivariate logistic regression analysis adjusting only for demographic data, MEOHP (OR = 3.26, 95% CI = 1.19-8.94, p = 0.029), MEHHP (OR = 3.98, 95% CI = 1.43-11.1, p = 0.016), MECPP (OR = 3.52, 95% CI = 1.01-12.2, p = 0.049), and MCiOP (OR = 4.55, 95% CI = 1.93-10.7, p = 0.005) were positively related to NAFLD defined by HSI and VCTE. The correlation strength varied with the concentration of phthalates, indicating a potential dose-response relationship. Adjusting for all covariates in multivariate logistic regression, only MCiOP (OR = 4.22, 95% CI = 1.10-16.2, p = 0.044), as an oxidative metabolite of DINP, remained significantly associated with NAFLD under the VCTE criterion, suggesting its potential role as a risk factor for NAFLD. This research highlights a significant association between DINP and NAFLD. These findings underscore the need for further investigation into the role of the phthalates substitutes in the pathogenesis of NAFLD and the importance of considering different diagnostic criteria in research.