AimTo examine the effects of employment engagement, classified by frailty and working status, on the incidence of disability in urban community‐dwelling older adults.MethodsWe used data from 6386 initially nondisabled residents aged 65–84 years from Ota City, Tokyo, Japan, in 2016. The observation duration was 3.6 years. Self‐administered questionnaires applied the Cox proportional hazard model by assessing the existence of frailty through Check‐List 15 (with a score ≥4 indicating the state of frailty), controlling for age, sex, living situation, education level, equivalent income, chronic conditions, body mass index, instrumental activities of daily living, self‐rated health, drinking and smoking status, and social activities. We evaluated the predictive value of working status (full‐time, part‐time, temporary, or nonworker) at baseline for cause‐specific disability (dementia‐type vs. non–dementia‐type) incidence, identified using the long‐term care insurance system's nationally unified database.ResultsOf the 6386 participants, 806 (63/1134 full‐time workers; 58/1001 part‐time workers; 61/547 temporary workers; 624/3704 nonworkers) presented with disabilities during the 3.6‐year‐long duration. Adjustments for conventional covariates showed that nonfrail full‐time and part‐time workers, as well as frail full‐time workers, had significantly lower risks of all‐cause disability incidence. Furthermore, nonfrail and frail full‐time workers had significantly lower risks of dementia‐type and nondementia‐type disabilities, respectively.ConclusionsThe incidence of disability in older adults was influenced by working and frailty status. Engaging in full‐time work thus prevents disabilities in older adults, regardless of their frailty status. Meanwhile, nonfrail older adults are able to avoid disabilities even when engaging in part‐time work. Geriatr Gerontol Int 2023; ••: ••–••.