The incidence of non-alcoholic fatty liver disease (NAFLD) in non-obese populations is also on the rise but is often overlooked. Therefore, this study aims to investigate the relationship between triglyceride-glucose (TyG) related parameters (TyG-BMI, TyG-WC, TyG-WHtR) and NAFLD in non-obese individuals to enhance early detection and intervention for NAFLD in this population. We conducted a cross-sectional analysis using data from the NAGALA database, covering 11,987 participants who underwent health examinations between 2004 and 2015. Logistic regression models were employed to assess the relationship between TyG-related parameters and NAFLD risk, incorporating cubic spline functions and smooth curve fitting to explore potential nonlinear relationships between TyG-related parameters and NAFLD risk. ROC curve analysis was performed to evaluate their predictive performance. After adjusting for confounding variables, the incidence of NAFLD in non-obese individuals increased with higher TyG-related parameters. Moreover, we identified nonlinear relationships between TyG index and its related parameters with NAFLD risk, along with critical inflection points. Additionally, the area under the ROC curve for TyG index and its related parameters were 0.7984, 0.8553, 0.8584, and 0.8353, respectively, with a more pronounced effect observed in the female population. There exists a nonlinear relationship between TyG-related parameters and the risk of NAFLD occurrence, revealing various turning points associated with NAFLD risk. Notably, TyG-related parameters outperform the TyG index in predicting NAFLD, especially in females. These findings establish crucial theoretical foundations for future research in NAFLD prevention and treatment for non-obese population.