This paper presents an algorithm for designing a system that classifies standard human facial expressions which are fear, disgust, sad , surprise, anger, happiness, and the normal expression . The facial expression that is presented in the input image of the system can be classified depending on extracting appearance features then, it is entered into neural network to complete the classification process using Matlab as a programming language. Multiple stages completed the work, which are, (collection images, pre-processing of the images, feature extraction, training neural network, classification and testing). Our system has been able to achieve the highest rating when the expression of anger reached 100 %, while the lowest rating was at the expression of sad by 30%.