To investigate the usefulness of percentage of free prostate specific antigen (FPSA/TPSA) in serum/PSA density [(F/T)/PSAD] in the diagnosis of prostate cancer.Two hundred and four patients who had been carried out transrectal ultrasound guided prostate biopsy, were involved in this study. Among them, 90 patients were proved to be suffering from prostate cancer, and other 114 patients were identified as benign prostate hypertrophy. The effect of total serum PSA level, FPSA/TPSA, PSAD and (F/T)/PSAD in the diagnosis of prostate cancer were investigated, and at the same time, selecting patients who should be carried out a prostate biopsy.The mean values of (F/T)/PSAD were significantly lower for patients with prostate cancer in different PSA levels (20.0 microg/L), when compared with benign prostate hypertrophy patients. This difference has arrived statistical significance (P < 0.05). (F/T)/PSAD could provide higher specificity for diagnosing prostate cancer than FPSA/TPSA or PSAD. Among all patients, at the same higher sensitivity (about 90%), the specificity of FPSA/TPSA, PSAD and (F/T)/PSAD was 31.6%, 45.6% and 64.0%, respectively. At the same time, it was suggested that clinicians use different cutoffs for (F/T)/PSAD in different PSA level. When PSA level of patients was no more than 4.0 microg/L, 2.5 as the commended cutoff for (F/T)/PSAD was preferred; if PSA level was between 4.0 microg/L and 20.0 microg/L, 0.8 was a more suitable cutoff; 0.5 also could be taken as an appropriate cutoff in case of PSA level being higher than 20.0 microg/L.Keeping high sensitivity, using of (F/T)/PSAD can improve the diagnostic specificity of prostate cancer significantly.