Research is an obligation for lecturers to develop their knowledge besides teaching. Until now all lecturer's research is managed by the Institute for Research and Community Service, often called LPPM. College research institute is a place where all lecturer research information can be obtained, problems encountered at tertiary research institutions are the grouping process in the field of lecturer research carried out by the LPPM data and information system section, this is also one of the problems related to text classification . Classification is the process of finding a model that describes and distinguishes classes or concepts that aim to be used to predict classes from objects whose class labels are unknown. Naive Bayes is a simple probabilistic based prediction technique based on the application of the Bayes theorem with strong independence assumptions. In this method there are deficiencies that can affect the accuracy caused by the Naive Bayes feature which is not always applicable. To deal with these problems, researchers conducted a feature selection process using Genetic Algorithms. The dataset used is 275 lecturer research data from all scientific fields. The results of experiments in this study indicate that the accuracy value increased by 26.06% with the use of Genetic Algorithms in the feature selection process.