Background: Reversing a neuromuscular blockade agent with sugammadex is known to lessen postoperative complications by reducing postoperative residual curarization. However, its effects on 90-day mortality are unknown. Therefore, this study aimed to compare the effects of sugammadex and neostigmine in terms of 90-day mortality after non-cardiac surgery. Methods: This retrospective cohort study analyzed the medical records of adult patients aged 18 years or older who underwent non-cardiac surgery at a single tertiary care hospital between 2011 and 2016. Propensity score matching and Cox regression analysis were used to investigate the effectiveness of sugammadex and neostigmine in lowering 90-day mortality after non-cardiac surgery. Results: A total of 65,702 patients were included in the analysis (mean age: 52.3 years, standard deviation: 15.7), and 23,532 of these patients (35.8%) received general surgery. After propensity score matching, 14,179 patients (3906 patients from the sugammadex group and 10,273 patients from the neostigmine group) were included in the final analysis. Cox regression analysis in the propensity score-matched cohort showed that the risk of 90-day mortality was 40% lower in the sugammadex group than in the neostigmine group (hazard ratio: 0.60, 95% confidence interval: 0.37, 0.98; P = 0.042). These results were similar in the multivariable Cox regression analysis of the entire cohort (hazard ratio: 0.62, 95% confidence interval: 0.39, 0.96; P = 0.036). Conclusions: This retrospective cohort study suggested that reversing rocuronium with sugammadex might be associated with lower 90-day mortality after non-cardiac surgery compared to neostigmine. However, since this study did not evaluate quantitative neuromuscular function in the postoperative period due to its retrospective design, the results should be interpreted carefully. Future prospective studies with quantitative neuromuscular monitoring in the postoperative period should be performed to confirm these results.