Diffusion-weighted imaging (DWI) is considered to be one of the dominant modalities used in prostate cancer (PCa) detection and the assessment of lesion aggressiveness, especially for peripheral zone (PZ) PCa. Computer-aided diagnosis (CAD), which is capable of automatically extracting and evaluating image features, can integrate multiple parameters and improve the detection of PCa. In this study, 13 quantitative image features were extracted from DWI by CAD, and diagnostic efficacy was analyzed in both the PZ and transition zone (TZ). The results demonstrated that there was a significant difference (P < 0.05) between PCa and non-PCa for nine of the 13 features in the PZ and five of the 13 features in the TZ. Besides, the prediction outcome of CAD had a strong correlation with the DWI scores that were graded by experienced radiologists according to the Prostate Imaging-Reporting and Data System Version 2 (PI-RADS v2).