In this paper we present an automatic segmentation of the Putamen shape from brain MRI based on wavelets and a neural network. Firstly we detect the Putamen region slice by slice using 1D wavelet feature extraction. Then fuzzy c-means technology is combined with edge detection to segment the objects inside the Putamen region. Finally features are extracted from the segmented objects and fed into a neural network classifier in order to identify the Putamen shape. Experiment shows the segmentation results to be accurate and efficient.