Objective: To develop and validate a prediction score for a successful retrograde procedure in chronic total occlusion (CTO) percutaneous coronary intervention (PCI). Methods: A total of 228 CTO lesions in 223 patients who underwent PCI by retrograde approach were analyzed. All subjects were randomly grouped to a derivation set and a validation set at a ratio of 2:1. A successful retrograde procedure was set as the end point. Each of the identified predictors for the end point by logistic regression was assigned 1 point and summed. Results: Independent predictors of a successful retrograde procedure were Werner's score [odds ratio (OR) 4.841, 95% confidence interval (CI) 1.952-12.005, p = 0.001], diameter of distal CTO segment (OR 5.263, 95% CI 2.067-13.398, p < 0.001) and tortuous collateral (type b; OR 0.119, 95% CI 0.032-0.444, p = 0.002). The predictive model developed in the derivation set stratified the difficulty of achieving a successful retrograde procedure into 4 grades - very difficult (10.5%), difficult (23.7%), intermediate (50.7%) and easy (15.1%) - and was demonstrated significantly in the validation set: very difficult (15.8%), difficult (18.4%), intermediate (47.4%) and easy (18.4%). The area under the receiver-operating characteristic curve was 0.832 ± 0.042 for the derivation set and 0.912 ± 0.041 for the validation set with an almost equal performance. Conclusions: According to the experience of our center, this model performed excellently in predicting the difficulty in achieving a successful retrograde procedure.