Gliomas, a type of brain tumor arising from glial cells, are a leading cause of pediatric cancer-related morbidity and mortality. They vary significantly in their aggressiveness, ranging from low grade gliomas to highly aggressive high-grade glioblastoma multiforme. Accurate grading of gliomas is crucial for effective treatment planning and improving patient outcomes. The aim of the study is to predict the glioma grading classification using a variety of machine learning classifiers on a publicly available glioma dataset with the incorporation of clinical and mutation features. The performance of these models was assessed through exploring all the possible combinations of methodological enhancements namely: feature selection step, class imbalance handling and application of hyperparameter tuning in the hope of potentially enhancing the predictive power of the classifiers. Results showed AdaBoost as the top performing model when all the enhancement methods were applied with a high accuracy of 85.71, 84.76% recall, 91.75% precision and 88.12% F1-score, equaling logistic regression best performance with a base configuration without applying any enhancement methods to the dataset. Overall, hyperparameter tuning is recommended as an effective approach to optimize the predictive performance of machine learning classifiers in glioma grading classification. The study’s findings could contribute to further the research and development of computer-aided diagnosis systems for the prediction of glioma grade, which can lead to enhanced treatment options for the patients.