Cyclic peptides are potentially therapeutic in clinical applications, due to their great stability and activity. Yet, designing and identifying potential cyclic peptide binders targeting specific targets remains a formidable challenge, entailing significant time and resources. In this study, we modified the powerful RFdiffusion model to allow the cyclic peptide structure identification (CycRFdiffusion) and integrated it with ProteinMPNN and HighFold to design binders for specific targets. This innovative approach, termed CycleDesigner, was followed by a series of scoring functions for efficient filtering. With the combination of effective cyclic peptide design and filtering, our study aims to further broaden the scope of cyclic peptide binder design.