Feature selection can improve classification accuracy and decrease the computational complexity of classification. Data features in intrusion detection systems (IDS) always present the problem of imbalanced classification in which some classifications only have a few instances while others have many instances. This imbalance can obviously limit classification efficiency, but few efforts have been made to address it. In this paper, a scheme for the many-objective problem is proposed for feature selection in IDS, which uses two strategies, namely, a special domination method and a predefined multiple targeted search, for population evolution. It can differentiate traffic not only between normal and abnormal but also by abnormality type. Based on our scheme, NSGA-III is used to obtain an adequate feature subset with good performance. An improved many-objective optimization algorithm (I-NSGA-III) is further proposed using a novel niche preservation procedure. It consists of a bias-selection process that selects the individual with the fewest selected features and a fit-selection process that selects the individual with the maximum sum weight of its objectives. Experimental results show that I-NSGA-III can alleviate the imbalance problem with higher classification accuracy for classes having fewer instances. Moreover, it can achieve both higher classification accuracy and lower computational complexity.