Leukemia is brought about by the quick generation of unusual white platelets. The high number of strange white platelets are not ready to battle contamination, and they impede the capacity of the bone marrow to create red platelets and platelets. Machine Learning techniques are widely used in the dignosis and classification of different leukemia types in the patients. In this paper, we have described the different machine learning algorithms like Support Vector Machines, k-Nearest Neighbour, Neural Networks, Naïve Bayes and Deep Learning algorithms which are used to classify leukemia into its sub-types and presented a comparative study of these algorithms.