BACKGROUND: Internet based self-help for panic disorder (PD) has proven to be effective. However, studies so far have focussed on treating a full-blown disorder. Panic symptoms that do not meet DSM-IV criteria are more prevalent than the full-blown disorder and patients with sub-clinical panic symptoms are at risk of developing PD. This study is a randomised controlled trial aimed to evaluate an Internet based self-help intervention for sub-clinical and mild PD compared to a waiting list control group. METHODS: Participants with mild or sub-clinical PD (N = 128) will be recruited in the general population. Severity of panic and anxiety symptoms are the primary outcome measures. Secondary outcomes include depressive symptoms, quality of life, loss of production and health care consumption. Assessments will take place on the Internet at baseline and three months after baseline. DISCUSSION: Results will indicate the effectiveness of Internet based self-help for sub-clinical and mild PD. Strengths of this design are the external validity and the fact that it is almost completely conducted online.