It has not been elucidated whether certain types of M1b prostate cancer (M1b PC) are associated with a poor outcome. The present study retrospectively identified predictive factors related to the outcome of M1b PC. The subjects were 104 patients who attended our hospital and received a diagnosis of M1b PC. The observation period ranged from 4 to 122 months (median, 43 months). The parameters investigated were: T classification, N classification, Gleason score (GS), pretreatment prostate-specific antigen (PSA) level, extent of disease (EOD) grade, alkaline phosphatase (ALP), lactate dehydrogenase (LDH), calcium, and hemoglobin (Hb) levels, platelet count, and the status of HER-2 overexpression as determined with a Hercep Test(TM) Kit using initial needle biopsy specimens for diagnosis. Log-rank test and Cox univariate analysis identified the following factors with statistically significant differences: pretreatment PSA ≥ 192, N1, GS ≥ 8, EOD grade 3+4, high LDH, high ALP, low Hb, and HER-2 overexpression. Multivariate Cox proportional hazard analysis identified the factors GS ≥ 8, high LDH, and HER-2 overexpression with significant differences. The hazard ratio was 5.962, 2.465, and 2.907, respectively, and the probability value was P=0.0218, P=0.0207 and P=0.0090, respectively. When the subjects with GS ≥ 8, high LDH, and HER-2 over-expression were classified as the high-risk group, the 5-year cause-specific survival rate was 51.2, 29.6, and 20.0%, respectively. The present study showed that M1b PC patients with GS ≥ 8, high LDH, and HER-2 overexpression have a very poor outcome and thus, should be treated as a high-risk group requiring close follow-up.