【Objective】 To identify the risk factors of patients of bone metastatic prostate cancer with high tumor load progressed to castration resistant prostate cancer (CRPC), establish a nomogram prediction model and evaluate its consistency and accuracy. 【Methods】 A total of 164 patients diagnosed by puncture and imaging during 2012 and 2022 were included.The general characteristics were analyzed with IBM SPSS software; the variables were screened with Cox regression; the multivariate risk factors with P<0.05 were included in the nomogram prediction model.The consistency and prediction accuracy of the model were evaluated with C-index, receiver operating characteristic (ROC) curve and calibration chart. 【Results】 In univariate analysis, initial prostate-specific antigen (PSA), prostate-specific antigen density (PSAD), Gleason score, T stage, alkaline phosphatase (ALP) and lactate dehydrogenase (LDH) were correlated with CRPC (P<0.05).Multivariate analysis showed that initial PSA, Gleason score, T stage, ALP and LDH were independent risk factors of CRPC (P<0.05).Based on the above five risk factors, a nomogram prediction model was constructed.The C-index was 0.801, the area under ROC curve (AUC) of 1-year progression-free survival (PFS) was 0.701 (0.608-0.794), and the AUC of 2-year PFS was 0.857 (0.767-0.947).The calibration chart showed that the prediction probability of the model was in good agreement with the actual probability. 【Conclusion】 Initial PSA, Gleason score, T stage, ALP and LDH are independent risk factors of CRPC.The predictive model may be an effective tool for the initial diagnosis of high tumor load bone metastatic prostate cancer, but more data are needed for internal and external validation.