Little is known about the association between serum alkaline phosphatase (ALP) levels and postoperative cognitive dysfunction (POCD) after general anesthesia. Thus, we investigated the association of serum ALP levels with POCD in patients who underwent surgery with general anesthesia in a retrospective cohort study. We retrospectively collected data from patients who underwent surgery with general anesthesia between May 2016 and June 2020. Serum ALP activity was detected using a p-nitrophenyl phosphate assay. Pre-and postoperative cognitive function were evaluated using the Chinese version of the Mini-Mental State Examination. Univariate and multivariate logistic regression were used to explore the effect of ALP on cognitive function. The incidence of POCD was 13.5%. Compared with the control group, the POCD group had higher ALP levels. The neuropsychological test results suggested that the scores of most items were lower in the POCD group than in the non-POCD group. Univariate logistic regression indicated that increased ALP levels were significantly associated with cognitive dysfunction (odds ratio = 1.15, 95% confidence interval: 1.13-1.18, P = .000). Multivariate regression showed that elevated ALP was still associated with POCD after adjusting for confounding factors (odds ratio = 1.16, 95% confidence interval: 1.13-1.18, P = .000). The spline regression model indicated the dose-response associations between ALP level and POCD risk (P for nonlinear trend < .001). Our study indicated that elevated serum ALP was an independent predictive factor of POCD at the 3-month follow-up. The occurrence of POCD could be associated with inflammatory status.