In this article we analyze the role of religion in the composition of Americans’ networks of anticipated emotional support. Drawing on data from the National Survey of Religion and Family Life, which contains information on multiple sources of potential emotional support, we use latent class analysis to uncover four different anticipated support profiles, which are organized along two dimensions of variation: religiosity and breadth. We label these profiles religious, secular, broad, and limited. Our analyses demonstrate associations between these anticipated support profiles and a person's gender, family status, age, race, socioeconomic status, and religious involvement. For instance, we find that Catholics are more likely than non‐Catholics to have secular rather than religious support profiles, and African Americans tend to have profiles that are either religious or limited. Finally, we show that these profiles have implications for well‐being. We contribute to research on religion and emotional support by describing how religious and secular sources combine into overall anticipated support profiles. Our conclusion addresses the implications of these findings for current scholarship on religion and emotional support networks.