Accurate diagnosis of prostate-cancer(PCa) remains challenging due to high false-negative rate of biopsy and low-specificity of the screening test. Computer-aided diagnosis(CAD) systems are increasingly being used for detection and diagnosis of PCa. Texture analysis has been proved to be a significant CAD tool in medical applications. The aim of this research was to investigate the role of texture parameters extracted from diffusion-weighted MRI and machine-learning classifiers in distinguishing PCa from normal peripheral-zone(PZ). The proposed methodology has achieved 93% accuracy using support-vector machine classifier. Experiments showed that the application of texture-analysis could improve the accuracy of identifying healthy and cancerous prostate-regions.