Blood, which has a very important role for human body transportation system, is mainly consist of three main elements: Red Blood Cells (RBCs), White Blood Cells (WBCs), and Platelets. Particular diseases could possibly attack these blood elements and leukemia or white blood cancer has been considered as one of the most fatal diseases that compromise white blood cells functions as the main immunity preservation agent. The numbers of new leukemia cases has been increased within these past years, while the survivability rate has not shown any improvement. The author suspect that one of the reasons is due to speed inadequacy in leukemia type classification and diagnostic process. In fact, in many cases rapid identification of the leukemia type are still troublesome, thus a simple, rapid, and reliable clinical decision support system is always become a necessity to support clinician diagnosis performance. According to Frecnch-American-British classification, there are four types of leukemia disease which is classified based on their severity level and infected cells type, they are Acute Myeloid Leukemia (AML), Acute Lymphoid Leukemia (ALL), Chronic Myeloid Leukemia (CML), and Chronic Lymphoid Leukemia (CLL). In this paper, the author would like to propose a novel approach to perform only acute leukemia type classification utilizing effective classifier architecture. The proposed method includes several conventional image pre-processing, image clustering, and image segmentation method to extract region of interest, along with feature extraction analysis using PCA method, and finally followed by neural network classifier which provide an excellent performance in classification process that reach 97.95% of accuracy to discriminate normal and cancerous cell images from given database