Although radical nephroureterectomy is the standard treatment method for upper urinary tract urothelial carcinoma, it is associated with a high risk of intravesical recurrence. There are no models for predicting IVR after RNU in patients with organ-confined UTUC. Therefore, we developed and validated a model for postoperative prediction of IVR after RNU. The development cohort consisted of 416 patients who underwent RNU with bladder cuff excision at our center between 1 January 2007 and 31 December 2015. Patient clinicopathologic data were recorded. Multivariate Cox proportional hazard ratio regression was used to build a predictive model with regression coefficients, backward step-wise selection was applied, and the likelihood ratio test with Akaike’s information criterion was used as the stopping rule. An independent cohort consisting of 152 consecutive patients from 1 January 2016 and 31 December 2017 was used for validation. The performance of this predictive model was assessed with respect to discrimination, calibration, and clinical usefulness. The predictors in this model included tumor stage, tumor diameter, tumor location, and tumor grade. In the validation cohort, the model showed good discrimination, with a concordance index of 0.689 (95% CI, 0.629 to 0.748) and good calibration. Decision curve analysis demonstrated that the model was also clinically useful. This study presents a good model that may facilitate individualized postoperative prediction of IVR after RNU in patients with organ-confined UTUC, and thus, may help improve postoperative strategies and facilitate treatment outcomes.