Within non-clinical samples the relationship between paranormal belief (PB) and well-being varies as a function of level of psychopathology. Accordingly, believers are best conceptualised as a heterogeneous set of sub-groups. The usefulness of previous findings has been restricted by conceptual methodological limitations. Specifically, overreliance on cross-sectional design, the assumption that believers constitute a homogeneous group, and consideration of direct effects. Acknowledging these limitations, the present study investigated whether profile membership derived from PB and psychopathology (schizotypy and manic-depressive experience) predicted well-being (i.e., stress, somatic complaints, life satisfaction and meaning in life) across time. Concurrently, analysis assessed the mediating effect of theoretically important variables (transliminality, happiness orientation, fearful and skeptical attitude). A sample of 1736 (Mage = 52, range = 18 to 88; 883 females, 845 males, eight non-binary) completed self-report measures indexing study constructs across time points. Latent profile analysis at baseline, identified three sub-groups varying in level of PB and psychopathology at baseline: Profile 1, moderate PB and high psychopathology; Profile 2, moderate PB and psychopathology; and Profile 3, moderate PB and low psychopathology. Path analysis demonstrated that Profile 1 (the highest psychopathology scoring profile) predicted higher negative and lower positive well-being over time in comparison with the other profiles. Moreover, Transliminality and Fearful Attitude positively mediated this relationship, whereas Skeptical Attitude produced negative mediation. These outcomes supported the presence of a sophisticated process underpinning the PB and well-being relationship. Overall, PB in the absence of psychopathology had no significant influence on well-being.