In the present review we detail the more universally accepted selection criteria in the various protocols of active surveillance in prostate cancer; we also identify and classify twenty nomograms/predictive models useful for decision making in active surveillance for prostate cancer. These models are classified in accordance to their prediction (High grade prostate cancer in radical prostatectomy specimen [Gleason grade > 7], understaging on biopsy compared to prostatectomy specimen, pathological stage, indolent cancer or progression after expectant therapy). We also detail the predictive variables used in each model for estimations, their internal validation parameters, the samples used to generate them, and the external validations if they were done. Many of them are presented with their URL address, where they may be consulted on line, this making easier their implementation. Finally we expose our thoughts about the use of probability density functions as a useful tool for validation of these predictive models that help in the definition of cut points facilitating their clinical use and credibility.