The association between inflammatory markers and depression has garnered increasing attention. The neutrophil percentage-to-albumin ratio (NPAR) is an emerging inflammatory marker, but its association with depression in middle-aged and elderly adults was not previously explored. The purpose of this study was to investigate the association through a national study in the United States. All study data were weighted to ensure representativeness. Multivariate logistic regression models were applied to explore the independent relationship of NPAR with depression in middle-aged and elderly adults. Restricted cubic splines were employed to examine potential nonlinear association, with turning points calculated using a recursive algorithm upon detecting nonlinearity. Stratified analyses and interaction tests were conducted to explore subgroup differences. In the model adjusted for all confounding factors, no significant relationship was found between NPAR and depression in middle-aged and elderly adults [1.02 (0.92, 1.12)]. Further sensitivity analysis indicated a potential U-shaped relationship between NPAR and depression in middle-aged and elderly adults, with the OR (95% CI) of 0.74 (0.60, 0.92), 0.87 (0.70, 1.08), 0.92 (0.72, 1.19) for Q2, Q3, and Q4, respectively, compared to Q1. The U-shaped association was confirmed by the restricted cubic spline. Subsequent analysis identified an inflection point at 14.05, revealing inverse relationships before and after this point. Subgroup analysis indicated sex differences in this association. This large-scale cross-sectional study identified a U-shaped association between NPAR and depression in American middle-aged and elderly adults.