Mechanisms of change represent the cornerstone of the therapeutic process. This study aimed to investigate how network models could be used to test mechanisms of change at a group level. A secondary aim was to investigate which of the several hypothesized mechanisms (emotion regulation, interoceptive awareness, and acceptance) are related to changes in psychological well-being. The sample comprised adult patients suffering from psychological disorders (N=444; 70% women) from 7 clinical sites in the Czech Republic who were undergoing groupbased multicomponent treatment composed mainly of psychodynamic psychotherapy (lasting from 4 to 12 weeks depending on the clinical site). Data were collected weekly using the multidimensional assessment of interoceptive awareness, emotion regulation skills questionnaire, chronic pain acceptance questionnaire-symptoms and outcome rating scale. A lag-1 longitudinal network model was employed for exploratory analysis of the panel data. The pruned final model demonstrated a satisfactory fit. Three networks were computed, i.e., temporal, contemporaneous, and between-person networks. The most central node was the modification of negative emotions. Mechanisms that were positively associated with well-being included modification, readiness to confront negative emotions, activity engagement, and trust in bodily signals. Acceptance of negative emotions showed a negative association with well-being. Moreover, noticing bodily sensations, not worrying, and self-regulation contributed indirectly to changes in well-being. In conclusion, the use of network methodology to model panel data helped generate novel hypotheses for future research and practice; for instance, well-being could be actively contributing to other mechanisms, not just a passive outcome.