Meadows, Randers, and Meadows developed a computer model, which studied relations between global resources, production, and consumption. Their model solves a simple system of ordinary differential equations under different scenarios and predicts the future of variables like population, food production, and various measures of pollution and use of resources. Due to the complexities of the observed system, such a simple model can not take into account all significant parameters and therefore cannot be used for accurate long term predictions. Therefore we developed a neural network based model, which we tested with historical data to reproduce the Meadows results. We also used this model to cluster countries based on similarity as measured by the observed variables in an attempt to predict, which characteristics are most significant for their development.