Early diagnosis of prostate cancer (PCa) as the second most common cancer in men is not associated with precise and specific results. Thus, alternate methods with high specificity and sensitivity are needed for accurate and timely detection of PCa. MicroRNAs regulate the molecular pathways involved in cancer by targeting multiple genes. The aberrant expression of the microRNAs has been reported in different cancer types including PCa. In this bioinformatics study, we studied differential expression profiles of microRNAs and their target genes in four PCa gene expression omnibus (GEO) databases. PCa diagnostic biomarker candidates were investigated using bioinformatics tools for analysis of gene expression data, microRNA target prediction, pathway and GO annotation, as well as ROC curves. The results of this study revealed significant changes in the expression of 14 microRNAs and 40 relevant target genes, which ultimately composed four combination panels (miR-375+96+663/miR-133b+143-3p+205/ C2ORF72+ENTPD5+GLYAT11/LAMB3+NTNG2+TSLP) as candidate biomarkers capable to distinguish between PCa tumor samples and normal prostate tissue samples. These biomarkers may be suggested for a more accurate early diagnosis of PCa patients along with current diagnostic tests.