Guided Internet-based cognitive-behavioral therapy (ICBT) has been found to be effective for social anxiety disorder (SAD) by several independent research groups. However, since the extent of clinically significant change demonstrated leaves room for improvement, new treatments should be developed and investigated. A novel treatment, which has generally been found to be effective, is cognitive bias modification (CBM). This study aims to evaluate the combination of CBM and ICBT. It is intended that two groups will be compared; one group randomized to receiving ICBT and CBM towards threat cues and one group receiving ICBT and control training. We hypothesize that the group receiving ICBT plus CBM will show superior treatment outcomes. Participants with SAD (N = 128), will be recruited from the general population. A composite score combining the scores obtained from three social anxiety questionnaires will serve as the primary outcome measure. Secondary measures include self-reported depression and quality of life. All treatments and assessments will be conducted via the Internet and measurement points will be baseline, Week 2, post-treatment, and 4 months post-treatment. There is no direct evidence of the effects of combining CBM and ICBT in SAD. Adding attention-training sessions to ICBT protocols could increase the proportion of participants who improve and recover through Internet-based self-help. ClinicalTrials.gov: NCT01570400.