Human brain tumor is a major health issue that may cause severe disabilities to a patient, ranging from loss of sensory feelings and paralysis to death. The accurate diagnosis of brain tumors that includes detection and localization is an essential procedure for effective treatment planning. Despite the fact that many brain tumor segmentation techniques have been developed in the past, effective tumor segmentation is still challenging because of the complex nature of the brain MRI image. It is difficult to differentiate between the tumor, which appears in the MRI image as a bright structure, and other bright structures that are not necessarily tumors. To address this problem, we propose a novel effective tumor segmentation method for brain MRI images. This technique is based on multiple threshold and object counting technique that can accurately detect the position of the tumor by being able to eliminate other similar structures. © 2018 IEEE.