We compared the performance of BCSA and GA for feature selection in this study. We used the Sonar Database from the UCI Machine in the experiments. The Sonar database has 60 attributes. Twenty-five features were specified using BCSA and 27 features were specified using GA. Classification accuracy of BCSA, GA, and all features were 99.04, 96.15 and 79.81, respectively. BCSA has achieved the most successful results in the study.