Segmentation involves dividing a brain Magnetic Resonance Image (MRI) image into its primary components, but a major challenge in this process is uncertainty, often stemming from factors such as noise and intensity non-uniformity. A comparison between two developed logics to enhance the MRI images has been done. A new model for brain MRI segmentation based on fuzzy inference system (FIS) and evidence theory has been presented. Dempster-Shafer fuzzy model (FDSIS) has used each rule as a witness in a FIS. The fuzzy Clustering Method (FCM) is an important mathematical structure in expressing knowledge and modelling complex nonlinear systems. In the presented method, the Fuzzy Clustering Method Integrated with Dempster-Shafer (FCMSUBET) belief structures are combined using Demaster's combination law. The results of the performance of the proposed method in brain MRI image using Dice and Tanimoto criteria were 0.924/0.859, 0.884/0.792, 0.933/0.874 for Cerebrospinal fluid (CSF), Gray Matter (GM), and White Matter (WM) respectively. While using the new proposed method the same results show better results for the same criteria used 0.929/0.867, 0.888/0.799, 0.93/0.87 respectively. The results show that the FCMSUBET method can overcome the shortcomings of the results of the FDSIS and that will give much better results of MRI brain segmentation.