The aim of the study was prediction of countries’ success that participating in the 2016 Olympic Games in Rio de Janeiro with intelligent method of MLP. This study was conducted on two qualitative (setting indices) and quantitative (collecting data from selected countries) steps. In the first phase, non-randomly and purposeful procedure, semi-structured qualitative and depth interviews was conducted with 28 elites that were aware of the issue of study up to theoretical saturation. After identification of indexes, Information of selected indicators (22 Indexes political, economic, social, cultural, technological and sporting as Theoretical Model (PEST+S)) was collected for 42 selected countries, in 40-years period, since 1976 Montreal Olympics to 2012 London. On the next step, from test of the conceptual model of multilayer perceptron networks (MLP) were used in comparison the actual values of Iran’s 22 indexes with predicted values in 2012. The results showed, in comparison the actual values with predicted values in 2012, the MLP has small errors in predicting of indices. Also, comparison of actual rank and predicted rank of 42 selected countries in 2012, MLP method had less Mean absolute error rate (0.4629). In final step, indices were estimated and rank of selected countries was predicted in Rio de Janeiro (2016). According to research results, Countries United States of America, China and the Great Britain will be in the first to third places, in 2016 Olympic Games. Also, Islamic Republic of Iran will be in 21st place among participating teams. Generally, using of Neural Networks Model, Iranian sport Policymaker’s can use identified indices to planning for successful participation in the Olympic Games.