Objectives Loneliness is seen as an important problem, contributing to serious health problems. As a baseline measurement for the evaluation of a community project aimed at reducing loneliness in Rotterdam, loneliness was measured, as well as potential correlates. This article describes models of social and emotional loneliness among older adults. Methods This study was conducted among 3,821 randomly selected community dwelling citizens of 65 years and older. Loneliness was measured by using the Jong-Gierveld loneliness questionnaire/scale. As potential correlates demographic, health and psychological and social variables were included. Data were subjected to multiple hierarchically regression analysis. Results Emotional loneliness was reported by 60% and social loneliness by 47% of the sample. Women were more emotionally lonely than men , while men reported more social loneliness than women. Emotional social support and quality of life were strongest in predicting emotional loneliness and social capital in the neighbourhood, companionship and instrumental support were strongest predictors for social loneliness. Demographic variables predicted variability in emotional loneliness. Conclusions The strength of the correlates differed between the two dimensions of loneliness. In the design of interventions to prevent and reduce loneliness among older adults, strategies should be developed aimed at the specific dimensions of loneliness.