Little is known on factors influencing cognitive decline in rural communities. Using the Atahualpa Project cohort, we aimed to assess whether the marital status or living arrangements influenced cognitive decline in community-dwelling adults living in an Ecuadorian rural village. The study included 629 Atahualpa residents aged ≥40 years who had a follow-up Montreal Cognitive Assessment (MoCA) repeated at least one year after baseline. Multivariate longitudinal linear models were fitted to assess differences between the marital status and living arrangements and cognitive decline (as the outcome). A total of 411 participants were married. The remaining 218 were either single, divorced or widowed (191 of them lived with family members and the others lived alone). Enrolled individuals contributed 2088.4 person-years of follow-up (mean: 3.3 ± 1.1 years). The mean baseline MoCA was 22.1 ± 4.5 points and the follow-up MoCA was 20.6 ± 4.8 points (p = 0.001). Overall, 394 (63%) individuals had lower MoCA scores at follow-up. A fully-adjusted longitudinal linear model showed no differences in MoCA decline across married and non-married individuals (β: −0.15; 95% C.I.: −0.55 – 0.26; p = 0.477). When the subset of non-married individuals was taken into account for analyses, multivariate longitudinal linear models showed no differences in the severity of cognitive decline across individuals living alone versus those living with family members (β: 0.28; 95% C.I.: −0.68 – 1.24; p = 0.572). In this study, the marital status or living arrangements had no influence on cognitive decline in the follow-up.