A system for face identification using facial features is designed. The system may identify e ither a complete unknown face image or a partial unknown face image, all these images represent the front view of the face. The data file of the system is built by storing each face image with its corresponding features that will be used in the identification process. In the identification of complete face images, the system is based on extracting the three most important parts (mouth, left eye, and right eye) which are represented as three subimages of size (32*32) and extracting the geometrical distances between various features of the face. The identification process involves three different levels of matching, each level uses a different method to perform the recognition task using the contents of the data file. In levell and Ievel2, the Walsh Power Spectrum method and the correlation method were chosen, respectively, to operate on the three subimages which are extracted from each face image in the system. The decision of the identification will depend on the computed coefficients at each level. In Ievel3 the Minimum Distance Rule was used as a matching procedure using geometrical features. In the identification of partial face images, the system is based on using the correlation method to find the best match between the unknown partial image and any of the face images in the data file. Experiments were carried out with a data file of 7 faces belonging to 5 different persons. An identification rate of 73.4% was achieved. The minimum requirements which are needed to run the proposed system are a computer of type IBM-AT or compatible, VGA monitor and a mouse.