Objectives: Evidence on the associations of childhood parental education with late-life cognitive function remains inconsistent. Additionally, studies focusing on dimensions of cognitive function are scarce, with none from China. Therefore, we examined the associations of childhood parental education with late-life cognitive function and it's dimensions and investigated potential interactions. Study design: A cross-sectional study. Methods: Participants were included from the third phase of Guangzhou Biobank Cohort Study. Mini-Mental State Examination (MMSE), Delayed 10-Word Recall Test (DWRT), and childhood parental literacy (could read or write) were collected. Multivariate linear and logistic regression were used. Results: Among 8891 participants aged >= 50 years, compared with those whose childhood parents could not read/ write, those with only the father, only the mother, or both parents able to read/write during childhood had higher scores on the MMSE and its dimensions, and lower odds of mild cognitive impairment (MCI). The coefficients ((3) and 95% confidence intervals (CI) for scores of MMSE were 1.25 (1.10, 1.41), 1.36 (0.86, 1.86), and 2.05 (1.89, 2.21) respectively; the odds ratios and 95% CIs for MCI were 0.38 (0.33, 0.45), 0.35 (0.18, 0.67), and 0.15 (0.12, 0.19) respectively. More pronounced effects were found in women, those aged >= 60, those with <= primary education, and in those whose childhood parents could read/write. Socioeconomic position mediated the association. Cognitive functions of attention and calculation in children of only fathers who could read/write ((3 = 0.43, 95%CI: 0.35, 0.50) and memory in children of only mothers who could read/write ((3 = 0.34, 95%CI: 0.16, 0.51) were associated with better outcomes. Consistent results were also found using scores from the DWRT. Conclusions: Childhood parental literacy was associated with late-life cognitive function and it's dimensions, which mediated by socioeconomic position. Specifically, having only a father or only a mother who could read/ write during childhood was associated with better cognitive function in attention/calculation and memory, respectively. These results add new evidence to support strategies for elderly cognitive health care.