Abstarct: In this thesis, a method for diagnosis of Attention-Deficit/Hyperactivity Disorder (ADHD) in children is presented. ADHD is a complicated multi-factorial and clinically heterogeneous disorder and it is common three times more frequent in child boys rather than child girls. It is estimated that about 5–10% of school-age children suffer from ADHD in worldwide. The exact cause of this disorder hasn’t been determined yet but advances in medical imaging of human brain and other studies in this field have proven the difference between child’s brain with this disorder and normal ones. Currently there is no effective drug to treat this disorder and best of them are strong narcotics and stimulus which have bad long-term side effects on human body. But by early diagnosis of the disease, the rates of drug usage and its advances could be controlled. In the first method, after preprocessing of images, proper features have been extracted from texture of them, and then we proceed to classified those features. Features extracted from the images, is in one dimension, because we have a collection of images that capture the resulting feature vector is large enough. When the images are in a dimension feature extraction, feature vector obtained from images of almost a third of the voxel in that case, we extract the feature vectors. In this method, detection is 70%. In the second and third method, After preprocessing of images using SPM toolbox, which is a powerful utility for preprocessing of MRI and fMRI images the LBP methods have been used for feature extraction. This is a powerful tool because of low computational complexity and can extract proper feature from brain images. After feature extraction, a method of feature reduction is performed. The difference between second and third method is due to using different classifier. For The second method SVM classifier and for the third one Neural network have been used for classification. Result of the second method for disorder(ADHD) diagnosis on four databaxses is between 81.4815 to 88.8889. The best Result of the third method for disorder(ADHD) diagnosis on four databaxses is between 94.1379 to 85.1852.