To develop nomograms predicting the incidence of castration-resistant prostate cancer (CRPC) and overall survival (OS) for de novo metastatic prostate cancer (PCa). Data from 449 patients with de novo metastatic PCa were retrospectively analysed. Patients were randomly divided into a training (n = 314, 70%) and a validation cohort (n = 135, 30%). Predictive factors were selected using a Cox proportional hazards model and were further used for building predictive models. The outcomes were incidence of CRPC and OS. Predictive factors included: Gleason score (GS), intraductal carcinoma of the prostate (IDC-P), Eastern Cooperative Oncology Group status, and alkaline phosphatase, haemoglobin and prostate-specific antigen levels. IDC-P and GS were the strongest prognosticators for both the incidence of CRPC and OS. Nomograms for predicting CRPC and OS had an internal validated concordance index of 0.762 and 0.723, respectively. Based on the β coefficients of the final model, risk classification systems were constructed. For those with favourable, intermediate and poor prognosis, the median time to CRPC was 62.6, 28.0 and 13.0 months (P < 0.001), respectively; and the median OS was not reached, 55.0 and 33.0 months, respectively (P < 0.001). We developed two novel nomograms to predict the incidence of CRPC and OS for patients with de novo metastatic PCa. These tools may assist in physician decision-making and the designing of clinical trials. © 2018 The Authors BJU International © 2018 BJU International Published by John Wiley & Sons Ltd.