Leukemia is cancer also known as blood cancer, is a deadly type of malignancy in humans that has the blood, bone marrow, and lymphatic tissues in its scope. This means that the sharp-eyed approach toward discovering leukemia is a better step toward heightening the chances of survival and the selection of an elk-suitable therapy is devised. Recent trends in the development of machine learning especially Support Vector machines SVM have shown great potential in medical imaging classification. This paper proposes a tkinter-based application that uses software that was classified by blood analysis using blood cells as an input. The feature extraction is accomplished using three approaches, i.e., fractal, run-length, and hybrid features which are then used to build SVM models. Evaluation models were done in terms of (accuracy, confusion matrices, and comparison with other similar studies). Fractal features were obtained in the SVM model at an accuracy rate of 0.70, run-length features at an accuracy of 0.86, and hybrid features achieved an accuracy 0.89 by combining feature extraction techniques. The paper contains a discussion of the conclusions made and, compares the results attained with the previous studies conducted in medical imaging to suggest ways in which future works in the field of leukemia detection may be improved.