Objective To investigate the role of vitamin D (VD) in improving cognitive dysfunction caused by sevoflurane anesthesia in aged mice Methods A total of 40 clean‑grade healthy 12‑month‑old C57BL/6J mice were utilized. According to the random number table method, they were divided into four groups (n=10): a control group (group C), a normal aged mice + sevoflurane group (group S), a VD(+) aged mice + sevoflurane group [group VD(+) ] and a VD(−) aged mice + sevoflurane group [group VD(−) ]. The VD(+) aged mice were established by gavage with osteotriol, while the VD(−) aged mice were constructed by feeding with special feeds, without light. Mice in groups S, VD(+), and VD(−) inhaled a mixture of 60% O2 and 3% sevoflurane daily in an airtight environment for three consecutive days, while those in group C inhaled 60% O2 daily for three consecutive days. On day 3 after successful modeling, the water maze test was conducted for six consecutive days. The escape latency (day 1 to day 5) and the number of platform crossings (day 6) were recorded for the four groups of mice. Mouse hippocampal tissues were taken, and the VD content was detected by enzyme‑linked immunosorbent assay, the activity of superoxide dismutase (SOD) was measured by hydroxylamine method, the activity of catalase (CAT) was measured by sodium molybdate colorimetry, the content of malondialdehyde (MDA) was measured by thiobarbituric acid method, the expression of synaptophysin (SYN) and post‑synaptic density‑95 (PSD‑95) were measured by Western blot. Results Compared with group C: group VD(+) showed increased VD content in hippocampal tissues of mice (P<0.05), while decreased VD content was seen in group VD(−) (P<0.05); there was no statistical difference in VD content in group S (P>0.05); group S and group VD(−) presented prolonged escape latency on days 2 to 5 (all P<0.05), with a reduced number of platform crossings (all P<0.05); group VD(+) showed a prolonged escape latency on day 2 (P<0.05), without statistical differences in escape latency and the number of platform crossings on days 3 to 5 (all P>0.05); the S, VD(+) and VD(−) groups presented increased MDA content (all P<0.05), decreased CAT and SOD activities (all P<0.05), and reduced SYN and PSD‑95 expression (both P<0.05). Compared with group S, group VD(+) showed shorten escape latency on days 3 to 5 (all P<0.05), with a increased number of platform crossings (P<0.05), reduced MDA content (all P<0.05), increased CAT and SOD activities (all P<0.05), and increased SYN and PSD‑95 expression (all P<0.05); group VD(−) showed prolonged escape latency on days 2 to 5 (all P<0.05), with a reduced number of platform crossings (P<0.05), increased MDA content (all P<0.05), decreased CAT and SOD activities (all P<0.05), and reduced SYN and PSD‑95 expression (all P<0.05). Compared with group VD(+), group VD(−) showed prolonged escape latency on days 2 to 5 (all P<0.05), with a reduced number of platform crossings (P<0.05), increased MDA content (all P<0.05), decreased CAT and SOD activities (all P<0.05), and reduced SYN and PSD‑95 expression (all P<0.05). Conclusions VD improves cognitive dysfunction in aged mice after sevoflurane anesthesia by reducing oxidative stress in the hippocampus and enhancing synaptic plasticity.