Background Formaldehyde is a common air pollutant in residential buildings, and the health risks caused by formaldehyde in residential buildings can not be ignored. Objective This study aims to evaluate the air concentration of formaldehyde in non-newly decorated houses in Ningbo and its possible health risks. Methods A total of 72 houses without any decoration in the past one year in Ningbo were selected by multi-stage random sampling method. From July 2018 to January 2019, the air samples of living rooms and bedrooms were collected and their temperature and humidity were also measured. The concentrations of formaldehyde were detected by AHMT method according to Standred method for hygienic examination of formaldehyde in air of residential areas — Spectrophotometric method (GB/T 16129—1995) , the health risk assessment model of U.S. Environmental Protection Agency was used to evaluate the non-carcinogenic risk and carcinogenic risk of formaldehyde, and Monte Carlo simulation was used for sensitivity analysis. Results The median (P₂₅, P₇₅) of formaldehyde concentration in the 72 houses was 0.019 (0.012,0.026) mg·m⁻³. Only one house showed a formaldehyde concentration that exceeded the national standard in the living room, and the total qualified rate of formaldehyde concentration was 98.61%. The median (P₂₅, P₇₅) of formaldehyde concentration in the bedroom was 0.019 (0.011, 0.031) mg·m⁻³, which was higher than that in the living room, 0.015 (0.010, 0.024) mg·m⁻³, and the difference was statistically significant. The median and 90th percentile of non-cancer risk (hazard quotient, HQ) of the 72 houses were 1.35 and 2.80, respectively, and the proportion of the houses with HQ>1 was 62.50%. The median and 90th percentile of cancer risk (CR) of the 72 houses were 1.12×10⁻⁴ and 2.32×10⁻⁴, respectively, and the proportions of the houses with CR>1×10⁻⁶, CR>1×10⁻⁵, and CR>1×10⁻⁴ were 100.00%, 100.00%, and 54.20%, respectively. After using Monte Carlo simulation, the median (90th percentile) of non-carcinogenic risk was reduced to 0.91 (1.94), where the median was lower than the national limit, and the proportion of samples with HQ>1 was 44.73%; the carcinogenic risk was reduced to 7.52×10⁻⁵ (1.79×10⁻⁴), and the proportions of samples with CR>1×10⁻⁶, CR>1×10⁻⁵, and CR>1×10⁻⁴ were 100.00%, 98.96%, and 34.37%, respectively. Conclusion The concentration of formaldehyde in non-newly decorated houses in Ningbo basically meets the national requirements, but it is still necessary to pay attention to the non-carcinogenic risk and carcinogenic risk caused by indoor formaldehyde, among which the carcinogenic risk is more important. Residents should prevent the harm of formaldehyde from its source by considering clean decoration materials and environmentally friendly furniture.