Introduction: Adolescents report using digital technologies for emotion regulation (digital ER), with the aim of feeling better (i.e., improving emotions and reducing loneliness). In this 7-day diary study, we investigated associations of digital ER, emotions, and loneliness, and tested whether prior emotional problems moderated these associations. Method: Participants were 312 Australian adolescents (Mage = 13.91, SD = 1.52; 44% boys). Daily surveys measured digital ER; end-of-day happiness, sadness, worry, anger, and loneliness; and peak sadness, worry, and anger. End-of day emotions were subtracted from peak emotions to calculate emotion recovery for sadness, worry, and anger. Participants were randomly selected from two symptom strata (high/low) defined by depression and social anxiety measures collected before the diary. Data were analyzed using multilevel path modeling. Cross-level interactions tested whether symptom strata moderated associations. Results: Digital ER was associated with more recovery from peak to end-of-day sadness and worry, but also with increased sadness, worry, anger, and loneliness by the next end-of-day. Higher end-of-day loneliness was associated with increased next-day digital ER. Prior emotional symptoms were not a significant moderator of daily digital ER and emotion associations. Conclusion: Adolescents who report more digital ER in a day show more recovery from the peak of negative emotion that day, but this recovery dissipates, with digital ER also associated with increased negative emotion and loneliness by the next day for all adolescents, regardless of prior symptom status. Lonelier adolescents use more digital ER by the next day, suggesting they need support to make social connections—online or offline.