Abstract-According to literature, a significant and up to date research direction to increase the performance level of automatic target recognition (ATR) systems is focused on the use of information coming from an appropriate set of EM sensors and high-quality decision fusion techniques, respectively. Consequently, in this paper a genetic optimized version of Sugeno's fuzzy integral is discussed. In addition, using a real database belonging to the high-resolution radar (HRR) imagery, the superiority of the proposed decision fusion technique related to its standard version and other well-known decision fusion methods is also demonstrated.