Background: Persons with chronic pain can be treated effectively with acceptance and commitment therapy (ACT). In this trial, we examined the effectiveness of guided and unguided ACT-based online treatment (ACTonPain) for chronic pain patients. Methods: 302 individuals were randomly assigned to ACTonPain with or without guidance (n = 100/101) or to a waiting-list control group (n= 101). The primary outcome was pain interference as measured by the Multidimensional Pain Inventory. The secondary outcomes were physical and emotional functioning, pain intensity, ACT process variables, quality of life, satisfaction with the intervention, adherence, and participants rating of overall improvement. The online measurements were carried out before randomization (T0) and 9 weeks and 6 months after randomization (T1 and T2, respectively). Intention-to-treat (ITT) data analysis was supplemented with additional per-protocol analyses. Results: The guided ACTonPain group showed significantly less pain interference than the control group in the ITT analysis (p = 0.01), with a moderate effect size at T1 and T2 (d = 0.58 respectively), corresponding to a number needed to treat (NNT) of 3.14 for both time points. Participants in the guided ACTonPain group also indicated higher pain acceptance (T1: d = 0.59; T2: d = 0.76). The unguided ACTonPain group showed to be significantly less depressed in comparison to the control group at at time T2 (d = 0.50). No significant differences with respect to effectiveness were found between the two ACTonPain groups (pamp;gt; 0.05). Conclusion: The online intervention ACTonPain is effective for persons with chronic pain when the program is guided. Further research in a variety of settings of health care is needed in order to determine whether and how ACTonPain can be implemented.