The strategy-situation fit hypothesis suggests that emotion regulation strategies are only beneficial to mental health if they meet contextual demands. Previous studies support this assumption but focused on properties of the emotional stimulus and reported cross-sectional associations with mental health, thus neglecting the social context and long-term mental health outcomes. To address these limitations, we examined (1) whether reappraisal, social sharing, and suppression varied depending on the social context (i.e., being alone, among close others, or nonclose others) and (2) whether specific strategy–context associations were linked to current, and (3) long-term mental health. The study consisted of two ecological momentary assessment (EMA) periods, separated by one year, and the second period occurred during COVID-19-related governmentally imposed social restrictions. This design allowed us to examine emotion regulation following social context changes. Our results indicate that emotion regulation varied by the social context. Reappraisal was used more frequently when being alone, suppression occurred more among nonclose others, and social sharing increased with close others. Regarding current mental health, more suppression was linked to higher depressive symptoms only when used in the presence of close others. In contrast, using suppression when being alone was linked to lower depressive symptoms. Finally, analyses with long-term outcomes revealed improved mental health when participants increased their use of reappraisal when being alone, and decreased reappraisal after a higher presence of close others. These findings could reflect the unique regulatory costs and benefits of different social situations and highlight the role of context-specific reappraisal for longer-term adaptiveness.