Introduction: Due to the prevalence of depression in diabetes and the importance of negative self-referrals (worry and rumination) in the severity of depression, this study aimed to evaluate the comparative effectiveness of acceptance and commitment-based therapy with and without compassion on worry, anger rumination, and self-critical rumination in non-clinical depressed diabetic Patients. Methods: The present study was a clinical trial with two intervention groups (first group: acceptance and commitment treatment without compassion intervention and second group: acceptance and commitment therapy with compassion) and a control group with two months follow-up. The statistical population consisted of all non-clinical depressed diabetic patients in the welfare centers of Mashhad in the spring of 2020. Using purposive sampling, 33 people were selected and then patients were divided into three groups and Interventions were assigned simple randomly (random sequence generation and allocation concealment). Data were collected by Sukhodolsky, Gotub & Cromwell Anger Rumination Questionnaire, Smart, Peter & Baer Self-Criticism Ruminant Questionnaire and Pennsylvania worry and analyzed by repeated measures analysis of variance in SPSS software version 25. Results: The results showed that in both intervention groups, the scores obtained on worry, rumination of anger and self-criticism in the post-test period decreased compared to the pre-test (p<0. 001). This effect has been stable over time (post-test and follow-up) (P> 0. 05). Conclusions: Acceptance and commitment-based therapy as an effective intervention can be used in medical centers to reduce negative self-referral processes (worry and rumination) of diabetic patients with depression vulnerability.