Weight loss was supposed to help with decreasing risk of premature mortality. However, results on this topic remain debatable and limited by study design. The present study aimed to investigate the association between weight loss and all-cause mortality among US adults with overweight or obesity in a national cohort study by using propensity score matching (PSM) analysis. A total of 5486 pairs of participants were matched in the National Health and Nutrition Examination Survey (NHANES, 2003-2015) after PSM. Hazard ratios (95% confidence intervals) (HRs (95% CIs)) were employed to evaluate the association between weight loss indicated by long-term weight loss (LTWL) and all-cause mortality by using Cox proportional hazards regression models. During a median follow-up of 6.8 years, 674 participants died from all-cause mortality. In each PSM match, compared with participants with LTWL < 5%, the HRs (95% CIs) for participants with LTWL of 5-9.9% (2877 pairs), 10-14.9% (1315 pairs), and ≥ 15% (1294 pairs) were 1.18 (0.83-1.68) (P = 0.366), 1.65 (1.17-2.34) (P = 0.005), and 1.91 (1.21-3.00) (P = 0.006), respectively. The significant increased risk of all-cause mortality for LTWL ≥ 15% remained among male, female, participants aged ≥ 65 years, without weight loss intention, with non-communicable diseases, and without exceeding estimated energy requirement. Weight loss especially for being ≥ 15% should be cautious for US adults with overweight or obesity.