Brain tumor analysis is very challenging task in Magnetic Resonance Images because images of brain are complicated, because complexity of the location, shape, size and its texture of tissues which has suffered damage through injury. Brain tumor is recognized by magnetic resonance imaging (MRI) consists of several stages. Segmentation is notable and necessary step in medical imaging for assortment and dissection. Performing the brain MR images segmentation it is very problematic work to find out brain tumor by manually. In present era MRI is very useful tool which helps for brain diagnosis. By Image processing we can able to detect this type of undesirable cells and upbraids the amount it spreads. The image from MRI scan will tell the presence of tumor in the brain, but we have to find the size of that tumor. In this paper, modified image segmentation techniques by applying MRI scan will be able detect the brain tumors. This paper will show how the image from MRI scan is adjusted to suitable contrast and tumor is separated from the original image and in this paper we present a technique for automatic brain tumor diagnostic system by magnetic resonance images. This system made up of three stages to identify & segment of a brain tumor. In very first stage, Magnetic resonance image of brain is received and pre-processing is done to evaluate the noise and also to enhance the image. And the second stage, edges are observed by implementing Gabor filter. In the last stage, Thresholding is the simplest method of image segmentation on the brain tumor. This segmentation increase the brightness of the image of that segment tumor and segmented image more operation by morphological and it remove the extra pixels segments. © 2016, International Journal of Pharmacy and Technology. All rights reserved.