Acceptance-based psychological interventions can potentially minimize the burden of chronic pain. This randomized controlled trial evaluated an internet-delivered, guided self-help intervention based on Acceptance and Commitment Therapy (ACT). A total of 238 chronic pain sufferers from the general population were randomly allocated to either ACT (n = 82), an internet-based control condition Expressive Writing (n = 79) or a waiting list condition (n = 77). Participants completed measures at baseline, posttreatment (3 months) and at a 3-month follow-up. At follow-up, ACT participants had improved in pain interference in daily life (primary outcome) compared to participants in Expressive Writing (Cohen's d = .47), but not compared to waiting list participants (p value = .11). Those who adhered to the ACT-intervention (48%) did improve significantly compared to waiting list participants (d = .49). ACT-participants also showed superior improvement on depression, pain intensity, psychological inflexibility and pain catastrophizing (d: .28-.60). Significant clinical improvement was present. Especially, 28% of ACT-participants showed general clinically relevant improvement in pain interference, as well as in pain intensity and depression (vs. Expressive Writing and waiting list 5%). Given these findings, internet-based ACT programs may be a promising treatment modality for chronic pain.