Objective: To analyze the cognitive mediation processes of an effective psychological group-based intervention promoting physical activity (based on motivational and volitional theories; MoVo-LISA) and to test for different underlying intervention mechanisms in the adoption and maintenance phase of physical activity behavior change. Method: In an RCT, N = 202 patients with coronary heart disease were randomly allocated to the group-based MoVo-LISA intervention or a control group (treatment-as-usual). Primary outcome was physical activity (in minutes per week, significant group differences: posttreatment: d = 1.03; 12-months follow-up: d = .57). Potential mediator variables were self-efficacy, outcome expectations, intention strength, self-concordance, action planning and barrier management. A mediation analysis of longitudinal data in SPSS AMOS v.24 was conducted. Results: The intervention was successful in modifying most of the examined cognitive variables in the adoption and/or the maintenance phase, with exception of self-concordance. Intention strength and self-concordance were significant predictors of behavior change in both phases. Findings suggest that the observed intervention effects on physical activity behavior were only partially mediated through the motivational variables self-efficacy and outcome expectation. Conclusions: The results contribute to opening “the black box” of processes behind an intervention and show that intervention effects of MoVo-LISA are partially mediated through the specified cognitive variables. The relevance of motivational variables (intention strength and self-concordance) in the adoption and maintenance phases is highlighted in this study and leads to relevant intervention design implications. Nevertheless, the differentiation of the two phases might help to improve intervention programs through a stronger focus on those cognitive parameters that are critical at a certain point in the behavior change process. (PsycInfo Database Record (c) 2022 APA, all rights reserved)