Social isolation and loneliness have been linked to adverse health outcomes such as depression in old age. However, limited data exist on the association of loneliness and social isolation with probable depression (PD) in low- and middle-income countries (LMICs), while psychosocial mediators are largely unknown. This study investigates the individual and joint associations of social isolation and loneliness with PD among older adults in Ghana. It quantifies the extent to which psychosocial factors mediate the associations. Cross-sectional data from the Aging, Health, Well-being, and Health-seeking Behaviour Study were analyzed. PD was defined as moderate to severe depressive symptoms with the Center for Epidemiologic Studies Depression (CES-D-9) scale. Loneliness and social isolation were assessed with the University of California, Los Angeles 3-item loneliness scale and the Berkman-Syme Social Network Index, respectively. Multivariable logistic models and PROCESS macro bootstrapping mediation analyses were performed. Among the 1,201 adults aged ≥50 years (Mage = 66.1 ± 11.9 years, 63.3% women), 29.5% PD cases were found. The prevalence of social isolation and loneliness was 27.3% and 17.7%, respectively. Loneliness (OR = 3.15, 95% CI = 3.26–5.28) and social isolation (OR = 1.24, 95% CI = 1.10–1.41) were independently associated with higher odds of PD. The loneliness and PD association was modified by spatial location (Pinteraction = 0.021); thus, the association was more pronounced in rural areas (OR = 7.06) than in urban areas (OR = 3.43). Psychosocial factors (e.g. sleep problems) mediated the loneliness/social isolation and PD association. Loneliness and social isolation were independently associated with a higher likelihood of PD, and psychosocial factors mediated the associations. Interventions to reduce PD in later life should also consider addressing loneliness and social isolation, as well as sleep problems.