Objective: The aim of the present study was to determine the efficacy of an Internet-based clinician-assisted computerized cognitive behavioural treatment (CaCCBT) programme for generalized anxiety disorder (the Worry programme). Methods: Forty-eight individuals meeting diagnostic criteria for generalized anxiety disorder (GAD) were randomly assigned to the Worry programme or to a waitlist control group. In the clinician-assisted Worry programme, participants complete six online lessons, weekly homework assignments, receive weekly email contact from a clinical psychologist, and contribute to a moderated online discussion forum with other participants. An intention-to-treat model was used for data analyses. The main outcome measures were Generalized Anxiety Disorder 7 Item Scale (GAD-7) and the Penn State Worry Questionnaire (PSWQ). Results: A total of 75% of treatment group participants completed all six lessons within the 9 week programme and post-treatment data were collected from 21/24 treatment group and 19/21 control group participants. Treatment group participants reported significantly reduced symptoms of worry as measured on the GAD-7 and PSWQ and reduced symptoms of depression as measured on the Patient Health Questionnaire 9 Item Scale (PHQ-9). Mean within- and between-groups effect sizes (Cohen's d) across the two measures of GAD were 1.3 and 1.1, respectively. Participants found the treatment programme acceptable and satisfactory. The clinician spent a total mean of 130 min per person over the programme. Conclusions: The Worry programme, the first randomized controlled trial of CaCCBT for GAD, resulted in clinically significant improvements. These results are consistent with literature indicating that Internet-based programmes, when combined with clinical guidance, can significantly reduce the symptoms of common mental disorders.