Objectives <p>Few studies have examined why Mindfulness-Based Stress Reduction (MBSR) programs are effective among teachers. The aim of this study was to examine possible mediators underlying the effectiveness of an MBSR program for Dutch elementary school teachers.</p> Methods <p>Data of a randomized controlled trial (RCT) with intervention (<i>n</i> = 72) versus control group (<i>n</i> = 74) were used. Questionnaires were completed before and after the intervention and at 3-month follow-up. First, 12 mediation models with mindfulness skills, self-compassion, teacher self-efficacy, or emotion regulation as mediator (M) at post-test and perceived stress, well-being, and classroom climate quality as dependent variables (Y) at 3-month follow-up were conducted. Second, when these factors have shown to be mediating factors in the simple mediation models, they were entered simultaneously in three multiple mediation models for each dependent variable.</p> Results <p>Self-compassion and emotion regulation mediated the effects of the MBSR program on perceived stress. Self-compassion mediated the effect of the MBSR program on well-being and teacher self-efficacy on classroom climate quality.</p> Conclusions <p>Findings imply that including intervention components targeting self-compassion and self-efficacy might increase the effectiveness of MBSR programs in reducing teachers’ perceived stress and improving teachers’ well-being and classroom climate quality. In addition, emotion regulation is an important aspect to maintain in current MBSR programs.</p> Preregistration <p>The study is preregistered (<a href="https://www.onderzoekmetmensen.nl/nl/trial/21817">https://www.onderzoekmetmensen.nl/nl/trial/21817</a>).</p>