BACKGROUND: The mechanism explaining how religiosity is linked to telomere length (TL) is unclear. The current study examines depression as a possible mediator. METHODS: In this cross-sectional study of 1,742 community-dwelling residents aged 55 or over, the Duke University Religion Index (DUREL) and Geriatric Depression Scale (GDS) were administrated during a routine health check. Peripheral blood leukocyte TL was determined using a q-PCR procedure. The Bootstrap methods PROCESS program was used to detect mediation. RESULTS: After controlling for sociodemographic variables, the religiosity was positively correlated with TL (p<0.05) and negatively correlated with depressive symptom (p<0.001). Depressive symptoms, in turn, was negatively correlated with TL (p<0.05) in the overall sample. Depressive symptoms significantly mediated the relationship between religiosity and TL (explaining 31.8% of the total variance) in the 65 years and older subgroup (p = 0.015). No significant mediation was found in the 55-64 age subgroup. LIMITATIONS: The cross-sectional design prevents making causal inferences. The non-random sampling method used in selecting participants may affect the external validity of the findings in terms of generalizing to Muslims throughout China or other religious groups. Potential mediators of the relationship between religiosity and TL and confounders such as physical health status, were not assessed. CONCLUSION: Religiosity was positively associated with TL in older mainland Chinese adults, and this association was partially mediated by depressive symptom in the 65 or older age group. This finding helps to explain why religiosity is related to cellular aging in older adults.