In this paper, a proposed hybrid algorithm using K-means and Fuzzy logic for brain segmentation, is developed, simulated and evaluated. The system identifies the white matter, gray matter and Cerebrospinal Fluid (CSF). The proposed system was tested using Magnetic Resonance Imaging (MRI), and evaluated in terms of the misclassification rate and percentage of clustering. The misclassification rate was found to be lesser in the proposed system as compared to the existing systems using K-means and Fuzzy logic. Further, the percentage of clustering is improved by the proposed system as compared to the existing algorithms. This work paves the way for future development of Neuro Fuzzy Kmeans algorithm in order to reduce the misclassification rate further in clustering the white matter, gray matter and CSF.