This paper presents an efficient approach to detect the Ischemic stroke condition from Magnetic Resonance Imaging (MRI) Diffusion images using bifurcation analysis of texture features with fuzzy C-means segmentation. The major scope of this research work is to identify the ischemic stroke in early stages to reduce the rate of mortality. Initially the input MRI slices are filtered to remove the multiplicative noises. Then the midline of the brain is traced for bifurcation analysis. Once the input image is bifurcated, texture based statistical features from both spatial and frequency domain is extracted. Finally the classification model is developed using Artificial Neural Networks. During classification, if the input image is detected as infarct, then fuzzy C-Means segmentation is carried out to localize the area of lesion. The proposed method is found to exhibit an accuracy of 94.43%. © 2015 by CESER PUBLICATIONS.