Exact measurements in brain diagnosis are difficult because of various shapes, area, and appearances of tumors. The tumor is an abnormal growth of body tissue; it can be cancerous or non-cancerous. There is a strong demand for automating the tumor detection and segmentation process. Thus, we required the computer aided diagnosis of brain tumor from MRI images to control the difficult problems in the manual segmentation. There are several methods available in the literature for medical image segmentation. In this paper, we introduced new segmentation method for detection of bias field image and classification of white matter and gray matter.From experimental results, BCFCM required less time than FCM algorithm.