Cognitive bias modification-interpretation (CBM-I) is a promising treatment modality for emotional disorders. But the moderate effect sizes emphasize the need to explore the potential mechanism of the CBM-I. Recent research has identified intolerance of uncertainty (IU) as a possible a trans-diagnostic risk and maintaining factor emotional disorders. The current study investigated the impact of a four-session IU-focused CBM-I on reduction of anxiety and determined whether IU mediated the relationship between IU interpretation bias and anxiety. 40 participants selected based on a high IU interpretation bias (IU-IB) were randomly assigned to an active (IU CBM-I) condition and a control CBM-I condition. By longitudinal design, anxiety and uncertainty questionnaires were administered before, after and one month after the intervention. Multilayer linear model analysis indicated that the active IU CBM-I was significantly associated with changes in anxiety from pre-to-post intervention as well as reductions in IU at post-intervention and month-one follow-up. The results also found that IU mediated the relationship between IU interpretation bias and anxiety.