Background This study examined the feasibility and preliminary effectiveness of d-cycloserine (DCS)-augmented cognitive behavioral therapy (CBT) for children and adolescents with difficult-to-treat Obsessive Compulsive Disorder, in a double-blind randomized controlled pilot trial (RCT). Methods Seventeen children and adolescents (aged 8-18 years) with a primary diagnosis of OCD, which was deemed difficult-to-treat, were randomly assigned to either nine sessions of CBT including five sessions of DCS-augmented exposure and response prevention (ERP) [ERP + DCS] or nine sessions of CBT including five sessions of placebo-augmented ERP [ERP + PBO]. Weight-dependent DCS or placebo doses (25 or 50 mg) were taken 1 hour before ERP sessions. Results At posttreatment, both groups showed significant improvements with 94% of the entire sample classified as responders. However, a greater improvement in the ERP + DCS relative to the ERP + PBO condition was observed at 1-month follow-up on clinician-rated obsessional severity and diagnostic severity, and parent ratings of OCD severity. There were no changes across time or condition from 1- to 3-month follow-up. Conclusions In this preliminary study, DCS-augmented ERP produced significant improvements in OCD severity from posttreatment to 1-month follow-up, relative to a placebo control condition, in severe and difficult-to-treat pediatric OCD. The significant effect on obsessional severity suggests that DCS augmentation might be associated with enhanced modification of obsessional thoughts during ERP, and warrants further investigation.