Hepatitis C is a liver disease whose infection is often silent and can lead to fibrosis or cirrhosis if it becomes chronic and goes undetected. It is generally spread through blood-to-blood contact. Hence, it is important to accurately classify blood donors as healthy blood donor or a person having Hepatitis C infection before blood transfusion happens. Nowadays, machine learning has been used in various domains including health care for accurate and fast results. This paper proposes a framework for accurate classification of blood donors using five machine learning algorithms, namely logistic regression, support vector machine, k-nearest neighbours, decision tree, and neural networks. The backward elimination technique is implemented for feature selection to improve the classification accuracy. The experimental results show that k-nearest neighbours perform better with the testing accuracy of 94.3% than other classifiers. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.