Neuropsychiatric symptoms (NPS) are prevalent across the dementia continuum and may emerge before overt cognitive impairment. Because many NPS involve affective and behavioral disturbances closely related to emotion regulation (ER), digital phenotyping using consumer-grade wearables offers an opportunity to identify digital biomarkers of ER from physiological and behavioral signals. Sixty-five middle-aged adults were monitored for eight weeks using wearables capturing sleep, activity, heart rate, heart rate variability, and respiration. Digital phenotypes were derived through hierarchical and k-medoids feature clustering, and their associations with ER outcomes were examined using regression models. Fourteen digital phenotypes spanning sleep, autonomic, activity, and respiratory domains were identified. Several phenotypes distinguished between different ER strategy-use profiles with 78% accuracy and were significantly associated with cognitive reappraisal, emotional clarity difficulties, and other ER outcomes, supporting their role as candidate digital biomarkers of emotion regulation. Overall, wearable-derived digital phenotypes capture meaningful individual differences in emotion regulation and highlight the potential of digital phenotyping to characterize emotional functioning in real-world settings. These findings support the development of scalable digital biomarkers relevant to brain health and dementia research.