| Objective To investigate the relationship between age, presence of blood in urine, tPSA, and total Gleason Score (GS) with bone metastasis of prostate cancer. Methods Clinical data of 277 patients with prostate cancer confirmed by pathology were retrospective analyzed. The patients underwent SPECT bone imaging and blood testing before treatment. Among them, 128 patients had bone metastasis and 149 patients did not. The differences of age, ALP, presence of blood urine, tPSA, total GS, primary GS, and secondary GS between the two groups were analyzed. The prognostic factors of bone metastasis were analyzed to obtain a mathematical model by Logistic regression analysis. The diagnostic efficiency of the mathematical model was evaluated by ROC curve. Results Except of age (P=0.736), there were significant differences in ALP, presence of blood in urine, tPSA, total GS, primary GS, and secondary GS between the two groups (P<0.01). Logistic regression analysis identified ALP, tPSA and primary GS as independent factors related to bone metastasis. The mathematical model based on ALP, tPSA and primary GS of prostate cancer patients in SPECT bone imaging was stated by P= ey/（1+ ey），y=-8.317+0.035×ALP+0.020×tPSA+0.912×primary GS. The AUC of ALP, tPSA , primary GS and mathematical model were 0.804, 0.808, 0.712 and 0.906, respectively. Conclusion ALP, tPSA and primary GS are the independent factors related to bone metastasis of prostate cancer. The mathematical model based on ALP, tPSA and primary GS of prostate cancer patients in SPECT bone imaging have high potential predicting bone metastasis.