This work presents an enhancement in accuracy for brain disorder detection using optimal unification. The strategy for detection of segments and brain regions causing medical conditions are described. This work demonstrates the application of multilateral filter and applied watershed method with EM-GM method. The most popular existing techniques of brain tumor detection are not optimal compared to this combination of Watershed and EM-GM technique with the proposed optimal unification technique. The result is optimally unified and achieved high accuracy. The multilateral filter enhances the image edges for better segmentation using signal amplitude moderation of the pixel. In the unification process, the optimal sets of segments are divided and finest merged results are considered with the brain regions detected with anomalies. Henceforth the number of possible medical investigations will be reduced.