Introduction: This is a study of the yearly dynamic of the HIV epidemic based on random walks has proven to be useful to take this highly variable phenomenon to a predictable behavior. Objective: Predict the behavior of the dynamic of the number of people living with HIV via a probabilistic random walk. Methodology: The yearly value of people living with HIV worldwide was analyzed from 1990 to 2009, based on probability spaces produced with a probabilistic random walk, and then, developed the prediction of the yearly value of people living with HIV for 2010, 2011 and 2012. Results: The yearly volume of people living with HIV was predicted with a 98.95% success rate in 2010, 98.82% in 2011 and 98.99% for the 2012 prediction. Conclusions: Mathematical orders were established based on the probabilistic random walk, establishing practically deterministic predictions of the number of people living with HIV which could be useful for public health decisions and to evaluate interventions.