AIM: To evaluate the visual outcomes of standard amblyopic treatment add-on training via perceptual learning in refractive amblyopic children and to identify the risk factors for treatment failure. METHODS: Retrospective charts were reviewed in children with refractive amblyopia who received standard treatment and add-on Cambridge Visual Stimulator (CAM) training. The add-on CAM group that was enrolled had worn full-corrected glasses for at least 2mo before training. A control group received only the standard treatment. Treatment success was defined as best-corrected visual acuity (BCVA) ≥20/25. The age, sex, initial BCVA, refractive errors, sessions and duration of training, and final BCVA were recorded. RESULTS: A total of 209 children (129 children in add-on CAM group and 80 children in control group) were enrolled. Seventy-six percent of unilateral and 87% of bilateral amblyopic children achieved treatment success. In children with unilateral or bilateral moderate amblyopia, the duration to reach BCVA ≥20/25 was significantly shorter in add-on CAM group than in control group. Poor initial BCVA (P<0.001) and high astigmatism (P=0.007) were risk factors for treatment failure after add-on CAM training. Age, sex, and types of refractive error were not associated with treatment success. CONCLUSION: Add-on CAM training is an effective strategy for visual improvement and can shorten the treatment course when the effect of standard treatment is limited in amblyopic children.