Background Severe health anxiety (hypochondriasis), or illness anxiety disorder according to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, is characterized by preoccupation with fear of suffering from a serious illness in spite of medical reassurance. It is a debilitating, prevalent disorder associated with increased health care utilization. Still, there is a lack of easily accessible specialized treatment for severe health anxiety. Objective The aims of this paper were to (1) describe the development and setup of a new internet-delivered acceptance and commitment therapy (iACT) program for patients with severe health anxiety using self-referral and a video-based assessment; and (2) examine the feasibility and potential clinical efficacy of iACT for severe health anxiety. Methods Self-referred patients (N=15) with severe health anxiety were diagnostically assessed by a video-based interview. They received 7 sessions of clinician-supported iACT comprising self-help texts, video clips, audio files, and worksheets over 12 weeks. Self-report questionnaires were obtained at baseline, post-treatment, and at 3-month follow-up. The primary outcome was Whiteley-7 Index (WI-7) measuring health anxiety severity. Depressive symptoms, health-related quality of life (HRQoL), life satisfaction, and psychological flexibility were also assessed. A within-group design was employed. Means, standard deviations, and effect sizes using the standardized response mean (SRM) were estimated. Post-treatment interviews were conducted to evaluate the patient experience of the usability and acceptability of the treatment setup and program. Results The self-referral and video-based assessments were well received. Most patients (12/15, 80%) completed the treatment, and only 1 (1/15, 7%) dropped out. Post-treatment (14/15, 93%) and 3-month follow-up (12/15, 80%) data were available for almost all patients. Paired t tests showed significant improvements on all outcome measures both at post-treatment and 3-month follow-up, except on one physical component subscale of HRQoL. Health anxiety symptoms decreased with 33.9 points at 3-month follow-up (95% CI 13.6-54.3, t11= 3.66, P=.004) with a large within-group effect size of 1.06 as measured by the SRM. Conclusions Treatment adherence and potential efficacy suggest that iACT may be a feasible treatment for health anxiety. The uncontrolled design and small sample size of the study limited the robustness of the findings. Therefore, the findings should be replicated in a randomized controlled trial. Potentially, iACT may increase availability and accessibility of specialized treatment for health anxiety. Trial Registration Danish Data Protection Agency, Central Denmark Region: 1-16-02-427-14; https://www.rm.dk/sundhed/faginfo/forskning/datatilsynet/ (Archived by Webcite at http://www.webcitation.org/6yDA7WovM)