Objective: To investigate the association between educational attainment and the onset of disability in activities of daily living (ADL), and to explore the mediating effect of social participation on such association. Methods: A longitudinal dataset was drawn from the China Health and Retirement Longitudinal Study (CHARLS, 2015 to 2018). The measurements of educational attainment, social participation and other covariates were identified from 2015, while the outcome measurement of ADL disability was constructed with data from survey 2018. Descriptive analyses were conducted, and basic characteristics and social engagement of the respondents were compared between illiterates and non-illite-rates using Chi-square test. Logistic regression was used to investigate the associations of educational attainment and social participation on the onset of ADL disability. Mediation analysis was employed to examine the mediator role of the social participation on the linkage from being illiterate to the ADL disability onset. Results: A total of 11 359 adults aged 45 years and above were included in the sample, of whom 3 222 were illiterates. The incidence of the onset of ADL disability of illiterates and non-illiterates were 10.4% and 6.2%, respectively. Among these respondents, only half of them were involved in social activities. Of all the 8 social activities, the percentage of interacting with friends (34.1%) was the highest, and the lowest percentage was observed in participating in an educational or training course (0.6%). Moreover, the percentages of participation in all these 8 social activities among illiterates were significantly lower than that of their educated counterparts (all P<0.001). The illiterate middle-aged and older adults were less likely to develop ADL disability in the follow-up period [adjusted odds ratio (aOR)=1.22, 95%CI: 1.02-1.45], and social participation was significantly associated with ADL disability onset (aOR=0.73, 95%CI: 0.63-0.85). Findings from mediation analysis illustrated that social participation accounted for 12.22% of the adverse effect of being illiterate on ADL disability onset. Conclusion: Social participation could buffer the negative effect of being illiterate on ADL disability onset in middle-aged and older adults, suggesting that engagement in social activities might have impact on prevention of impairments in physical function, especially for middle-aged and older illiterates.