Internet made a big revolution in the real world and thus poses so many challenges to the researchers by generating an enormous amount of data. The data generated contains an enormous amount of unwanted information. Before processing with such a dataset, the important features present in the dataset must be retrieved. The feature selection process is important because the performance of a model built for the purpose of classification, prediction or clustering depends mainly on the number of relevant features present in the dataset. In this proposed work, the real coded genetic algorithm is used to find the important features by considering the fuzzy rough degree of dependency as its fitness function for finding out optimum features for agricultural dataset, iris dataset and Pima Indian diabetes dataset. The experimental results show that the proposed work produces relevant features by maintaining classification accuracy. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021