Background: The marker of proliferation Ki-67 (KI67) is a well-known biomarker reflecting cell proliferation activity and was reported to be associated with the treatment efficacy of chemotherapy in different tumors. However, it is unknown whether KI67 also has a role in predicting the efficacy of the next-generation hormone therapies, e.g., abiraterone, for patients with prostate cancer (PCa). Methods: Clinicopathological data of 144 men with metastatic PCa who received abiraterone therapy were retrospectively collected. The KI67 positivity was examined by immunohistochemistry (IHC) using the prostate biopsy specimen. The predictive value of KI67 and other factors on the therapeutic efficacy of abiraterone treatment was explored. Kaplan-Meier curve and COX regression analysis were used for survival analysis. The endpoints were PSA progression-free survival (PSA-PFS) and radiographic progression-free survival (rPFS) according to the PCWG3 criteria. Results: Among the 144 included patients, 82 (56.9%) and 62 (43.1%), respectively, received abiraterone therapy at metastatic hormone-sensitive (mHSPC) and castration-resistant PCa (mCRPC) stage. The median KI67 positivity was 20% for the total cohort (interquartile range: 10%-30%). When taken as a continuous variable, KI67 positivity was adversely related to both PSA-PFS (HR, 95%CI: 1.02, 1.01-1.03, P=0.001) and rPFS (HR, 95%CI: 1.02, 1.02-1.03, P=0.001) of the abiraterone therapy, while was not associated with PSA response. The results of multivariate COX regression analysis implied that, after adjusting by mHSPC/mCRPC stage, ISUP grading, pain score, visceral metastasis, and metastatic burden, KI67 was still an independent predictor of PSA-PFS (HR, 95%CI: 1.02, 1.01-1.03, P=0.001) and rPFS (HR, 95%CI: 1.02, 1.01-1.03, P=0.008) of the abiraterone therapy. Moreover, when patients were divided into two groups with high and low KI67, using cutoffs of 10%, 20%, and 30%, KI67 remained a significant predictor of both PSA-PFS (HR: 1.67, 1.91, 2.29; P value: 0.052, 0.004, <0.001) and rPFS (HR: 1.54, 1.83, 1.88; P value: 0.119, 0.008, 0.008) of the abiraterone treatment. Subgroup analysis based on whether patients received abiraterone at mHSPC or mCRPC stage suggested that KI67 30% (median PSA-PFS: 11.4- vs. 27.6-Mo, P<0.001; median rPFS: 16.6- vs. 33.6-Mo, P=0.001) was the optimal cutoff to maximize patient prognostic differentiation for the mHSPC cohort, while KI67 20% (median PSA-PFS: 6.9- vs. 14.9-Mo, P=0.033; median rPFS: 10.1- vs. 16.5-Mo, P=0.030) was the optimal cutoff for the mCRPC cohort. Conclusions: Our study revealed that KI67 positivity in prostate biopsy specimens was a robust predictor of abiraterone efficacy for patients with advanced PCa. Further validation using datasets from other centers is needed to strengthen the findings of our work.