Cellular abnormality leads to brain tumor development. It is one of the main causes of mortality worldwide for adults and children. Early tumor discovery can avert millions of mortalities. Magnetic Resonance Imaging (MRI) is one of the most popular imaging techniques that can be used for the earlier detection of brain tumors, so that may improve the survival of patients. Tumor visibility is improved in MRI, which facilitates subsequent treatment. This research tries to detect brain tumors early on. The suggested CAD system that uses MRIs has the potential to help doctors and other specialists to find the existence of brain tumors. This work makes use of machine learning to enhance classification accuracy. This work is carried out in many sequential steps that include preprocessing using the median filter for MRIs noise removal, feature extraction using Grey Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP) for extracting the features of tumors, then Principal Component Analysis (PCA) and Information Gain (IG) are used as Feature selection algorithms, finally, the classification step is performed using different types of machine learning classifiers to determine and classify the MRIs as tumorous or non-tumorous. The experimental results of the proposed method, which uses a combined feature vector of GLCM and LBP features, show 98% accuracy using IG and 97% accuracy using PCA.