BackgroundInternal motivation and good psychological capabilities are important factors in successful eating-related behavior change. Thus, we investigated whether general acceptance and commitment therapy (ACT) affects reported eating behavior and diet quality and whether baseline perceived stress moderates the intervention effects.MethodsSecondary analysis of unblinded randomized controlled trial in three Finnish cities. Working-aged adults with psychological distress and overweight or obesity in three parallel groups: (1) ACT-based Face-to-face (n = 70; six group sessions led by a psychologist), (2) ACT-based Mobile (n = 78; one group session and mobile app), and (3) Control (n = 71; only the measurements). At baseline, the participants’ (n = 219, 85% females) mean body mass index was 31.3 kg/m2 (SD = 2.9), and mean age was 49.5 years (SD = 7.4). The measurements conducted before the 8-week intervention period (baseline), 10 weeks after the baseline (post-intervention), and 36 weeks after the baseline (follow-up) included clinical measurements, questionnaires of eating behavior (IES-1, TFEQ-R18, HTAS, ecSI 2.0, REBS), diet quality (IDQ), alcohol consumption (AUDIT-C), perceived stress (PSS), and 48-h dietary recall. Hierarchical linear modeling (Wald test) was used to analyze the differences in changes between groups.ResultsGroup x time interactions showed that the subcomponent of intuitive eating (IES-1), i.e., Eating for physical rather than emotional reasons, increased in both ACT-based groups (p = .019); the subcomponent of TFEQ-R18, i.e., Uncontrolled eating, decreased in the Face-to-face group (p = .020); the subcomponent of health and taste attitudes (HTAS), i.e., Using food as a reward, decreased in the Mobile group (p = .048); and both subcomponent of eating competence (ecSI 2.0), i.e., Food acceptance (p = .048), and two subcomponents of regulation of eating behavior (REBS), i.e., Integrated and Identified regulation (p = .003, p = .023, respectively), increased in the Face-to-face group. Baseline perceived stress did not moderate effects on these particular features of eating behavior from baseline to follow-up. No statistically significant effects were found for dietary measures.ConclusionsACT-based interventions, delivered in group sessions or by mobile app, showed beneficial effects on reported eating behavior. Beneficial effects on eating behavior were, however, not accompanied by parallel changes in diet, which suggests that ACT-based interventions should include nutritional counseling if changes in diet are targeted.Trial registrationClinicalTrials.gov (NCT01738256), registered 17 August, 2012.Electronic supplementary materialThe online version of this article (10.1186/s12966-018-0654-8) contains supplementary material, which is available to authorized users.