AbstractWildfires cause major damage and losses around the world. Such damages range from human and economical losses to environmental ones. Therefore, having models to predict their behavior can be a key element in the process of firefighting. In this paper, we present a comparative study between two methods we have developed. Both methods use Statistical Analysis, Parallel Evolutionary Algorithms and High Performance Computing, respectively named: Evolutionary-Statistical System (ESS) and Evolutionary-Statistical System with Island Model (ESS-IM). In this study, we have compared these two methods in terms of quality of prediction and performance in the parallel environment.