Purpose: To analyze variability, associated factors, and the design of nomograms for individualized testosterone recovery after androgen deprivation therapy (ADT) withdrawal.Methods: A longitudinal study was performed on 208 patients in 2003-2019 period. The castrate and normogonadic levels were defined as testosterone, 0.50 and 3.50 ng/ml respectively. Cumulative incidence curve describes testosterone recovery. A univariate and multivariate analysis was performed to predict testosterone recovery with the candidate prognostic factors: PSA at diagnosis, Clinical stage, biopsy Gleason score, age at cessation of ADT, duration of ADT, primary therapy for patients, and LHRH agonist. Results: The median followup of the study was 80 months, interquartile range (49,99). The 25% and 81% of patients did not recover the castrate and normogonadic level, respectively. Months of ADT and age at ADT withdrawal were significant predictors for testosterone recovery. We built two nomograms of testosterone estimation recovery at 12, 24, 36 and 60 months. The castration recovery model shows good calibration. The c-index was 0.677, with areas under the ROCcurve (AUC) of 0.74, 0.78, 0.78 and 0.78, at 12, 24, 36 and 60 months, respectively. The normogonadic recovery model had an overestimation of high probabilities. The cindex was 0.683, with AUC values of 0.81, 0.71, 0.71 and 0.70 at 12, 24, 36 and 60 months, respectively.Conclusion: Depending on the age of patients and time of treatment, clinicians can discontinue ADT to maintain castrate levels without treatment with enough confidence, or even recover testosterone to normogonadic levels in short courses of treatment with high probabilities.