OBJECTIVES: This study aimed to determine if mindfulness-based cognitive therapy (MBCT) engenders improvement in headache outcomes via the mechanisms specified by theory: (1) change in psychological process, (i.e., pain acceptance); and concurrently (2) change in cognitive content, (i.e., pain catastrophizing; headache management self-efficacy). DESIGN: A secondary analysis of a randomized controlled trial comparing MBCT to a medical treatment as usual, delayed treatment (DT) control was conducted. Participants were individuals with headache pain who completed MBCT or DT (N=24) at the Kilgo Headache Clinic or psychology clinic. Standardized measures of the primary outcome (pain interference) and proposed mediators were administered at pre- and post-treatment; change scores were calculated. Bootstrap mediation models were conducted. RESULTS: Pain acceptance emerged as a significant mediator of the group-interference relation (p<.05). Mediation models examining acceptance subscales showed nuances in this effect, with activity engagement emerging as a significant mediator (p<.05), but pain willingness not meeting criteria for mediation due to a non-significant pathway from the mediator to outcome. Criteria for mediation was also not met for the catastrophizing or self-efficacy models as neither of these variables significantly predicted pain interference. CONCLUSIONS: Pain acceptance, and specifically engagement in valued activities despite pain, may be a key mechanism underlying improvement in pain outcome during a MBCT for headache pain intervention. The theorized mediating role of cognitive content factors was not supported in this preliminary study. A large, definitive trial is warranted to replicate and extend the findings in order to streamline and optimize MBCT for headache.