Often, pathological Gleason Score (GS) and stage of prostate cancer (PCa) were inconsistent with biopsy GS and clinical stage. However, there were no widely accepted methods predicting upgrading and upstaging PCa. In our study, we investigated the association between serum testosterone and upgrading or upstaging of PCa after radical prostatectomy (RP). We enrolled 167 patients with PCa with biopsy GS ≤6, clinical stage ≤T2c, and prostate-specific antigen (PSA) <10 ng ml-1 from April 2009 to April 2015. Data including age, body mass index, preoperative PSA level, comorbidity, clinical presentation, and preoperative serum total testosterone level were collected. Upgrading occurred in 62 (37.1%) patients, and upstaging occurred in 73 (43.7%) patients. Preoperative testosterone was lower in the upgrading than nonupgrading group (3.72 vs 4.56, P< 0.01). Patients in the upstaging group had lower preoperative testosterone than those in the nonupstaging group (3.84 vs 4.57, P= 0.01). In multivariate logistic regression analysis, as both continuous and categorical variables, low serum testosterone was confirmed to be an independent predictor of pathological upgrading (P = 0.01 and P= 0.01) and upstaging (P = 0.01 and P = 0.02) after RP. We suggest that low serum testosterone (<3 ng ml-1 ) is associated with a high rate of upgrading and upstaging after RP. It is better for surgeons to ensure close monitoring of PSA levels and imaging examination when selecting non-RP treatment, to be cautious in proceeding with nerve-sparing surgery, and to be enthusiastic in performing extended lymph node dissection when selecting RP treatment for patients with low serum testosterone.