Abstract: Image processing acts as a vital part of distinct healthcare applications to assist the computer based disease diagnosis models. The brain tumor is found to be a lifethreatening kind of cancer and the earlier identification results in improved survival rate. Magnetic Resonance Image (MRI) is the commonly available imaging technique used for recording the glioma for the medical examination. Brain tumor segmentation and classification processes aim to separate the tumor from the normal brain tissues and identify the presence of brain tumor or node, which is helpful for further treatment. But it remains a crucial process owing to the irregular form and unclear boundaries of tumors. This paper presents a novel Optimized Kapur’s Entropy with adaptive neuro fuzzy inference system (OKE-ANFIS) model for brain tumor segmentation and classification using MRI images. The presented model involves skull stripping based preprocessing, segmentation, and classification. The presented model employs OKE based segmentation technique where the optimal choice of threshold values is decided using moth flame optimization (MFO) algorithm. Besides, histogram of gradients (HOG) based feature extractor is employed. In addition, ANFIS based classifier is utilized to determine the class label of the applied MRI images. For experimental validation of the OKE-ANFIS model, a wide range of simulations was performed and the results are investigated under several aspects. The experimental outcome ensured the superiority of the OKE-ANFIS model.